CRYSTAL RADIO:

WjMpP””
FUNDAMENTALS

Pickard files Mass-produced j

"catwhisker" detector tube sets overtake

r l: patent
1911
the crystal
1925
i j
i: I
1906 1922
—Histone v/d Radiqnwoojy app[ies Manufactured
__________ undum as detector crystal sets available
 5

'

:
i
i

I«

n
f
!
Crystal Radio:
History, Fundamentals, and Design
 '

!
r

i
Crystal Radio:
History, Fundamentals, and Design
P. A. Kinzie

V
The Xtal Set Society
MWV.mldnlghteclcr\CC.COtt\
1-^00-327-1771
Copyright © 1996 Phil A. Kinzie
All Rights Reserved.

Printed in Lawrence, Kansas

ISBN: 1-887736-05-0
CONTENTS
Chapter 1 The History of the Crystal Set 7
Introduction 7
The Discovery of Solid-State
Rectification 7
Solid-State Detection: the Efforts
that Failed or Were Forgotten 10
Solid-State Detection: the Prelude
to Success 14
The First Practical Crystal Detectors 17
Continuing Development 24
The Competitive Years, and the Role
of the Crystal Detector 27
A New Industry 32
Survival as a Specialty 34
Mass-Produced Rectifiers 39
Bibliography 42

Chapter 2 Fundamentals of Receiver Systems 45

Some Comments on Definitions 45
Signal Reception Requirements 47
Antennas 48
Ground 54
Lightning Protection 56
Signal Selection 58
Antenna Tuning 59
Detector Circuit Coupling and Tuning 63
Tuned Transfer Circuits 66
Wave Traps 69
Signal Detection 72
The Detection Process 73
The Audio Output 78
Bibliography 81
Chapter 3 Crystal Set Design 83
Introduction 83
Design Criteria 84
Sensitivity and Selectivity 85
Size 90
Simplicity 91
Appearance 92
Antenna Compatibility 93
Design Options for Applications 94
Single Tuned Receivers 95
Receivers with Two Tuned Circuits 99
Receivers with Three Tuned Circuits 100
Variable Inductance Tuned Receivers 103
Wave Trap Applications 103
Bibliography 105

Chapter 4 Crystals for Detectors 107

Materials for Detector Applications 107

Contacting Methods 110
Application Technique 114
Bibliography 118

About the Author 119

The Xtal Set Society 121
Other Crystal Set Books 122
VA CHAPTER 1
THE HISTORY OF THE CRYSTAL SET

Introduction

The crystal set came into existence because of the

invention of just one component, the detector.
Everything else was easily adapted from the parts of existing
circuitry. It was the development of this one key item that
made the crystal set a reality. Therefore the history of the
crystal detector itself is important in order to cover crystal set
development. In the following sections first we cover the
discovery of solid-state rectification, then the early efforts to
utilize that discovery, the first practical applications, and
finally, the evolution of both detector and receiver up to
today's mass-produced solid-state diode.

This process took time. Some 70 years elapsed between the

first successful laboratory work demonstrating the property
of rectification in solids, and the first practical solid-state di­
odes. Surprisingly enough, relatively few people were
involved to any great degree over much of this period. It was
not until about the time of World War II that large scale ap­
plied research and development took place. All of these
aspects of the development process are covered in the
following sections.

The Discovery of Solid-State Rectification

The latter half of the 19th century was a time when scientific
discoveries were made at a rapid rate, and in many cases
laboratory results accumulated far faster than successful
theories to explain them. This was the situation with studies
of electrical conduction in solids, where theory and

7
 oriment had established only two methods of conduction,
experiment.
by Ohm's Law or by electrolysis.

A young German physicist named Ferdinand Braun was the

first investigator to present unexpected but conclusive ex­
perimental results to the contrary. These proved that another
form of conduction exists, and that the property of rectifica­
tion, sometimes called unilateral conductivity, can be
demonstrated in a number of natural and artificially produced
solids. This experimental work was done many years before
there was adequate theory to explain the results.

Braun was bom in Fulda, a town in what is now Germany,

on June 6, 1850. His aptitude for experimental physics and
his teaching skills were already evident by the time he was in
his twenties, and he later became a Professor of Physics. In
the early 1870's Braun began investigating electrical con­
duction in certain solids that appeared to have interesting
properties. He chose to test a group of natural and laboratory
prepared metal sulfides, and he observed that current flow in
a number of these compounds obeyed neither Ohm's Law nor
the law of electrolysis.

Braun's method of mounting his specimens was unique for

that time, and it turned out to be the key factor in producing
his experimental results. His procedure consisted of first
placing the irregularly shaped sample in a ring of silver wire.
Then a second silver wire, crimped to produce spring action,
was installed to press down on the top of the sample by con­
tact at its pointed end. The result was a point contact
junction, which we now know is a requirement for efficient
rectification of current in the case of a large class of sub­
stances. Earlier experimenters had clamped their specimens
between metal plates, or had used some similar method

8
which exposed relatively large areas on each side of the
sample to conductive surfaces. With that technique and with
the particular substances tested, the results had been incon­
clusive at best.

Braun's results were quite unexpected; specimens connected

by the new method often conducted electricity much better in
one current flow direction than the other. He published his
initial results in 1874, which were at first controversial.
Some of Braun's colleagues disagreed with him while others
reported similar results. Braun responded with three more
papers on the subject, the last one being nine years later in
1883. By that time the phenomenon was fully established;
there were many crystalline substances known to act as rec­
tifiers, having the ability to conduct better in one direction
than in the other.

After this period, interest in crystal rectifiers faded. A suc­

cessful theoretical explanation of unilateral conductivity was
years away in the far future, so there was no incentive to plan
further experiments, much less a hint of what direction to
take. In addition, there was not even a clue of a practical
application of the effect. Finally, there were so many prom­
ising developments taking place in other fields that
investigators quite naturally turned elsewhere.

Braun's final paper preceded the demonstration of electro­

magnetic wave propagation in space by Heinrich Hertz in
1888, and was years ahead of Guglielmo Marconi's success­
ful wireless transmission experiments in 1895. Therefore, no
thought was given by anyone about the possibility of using a
crystal rectifier as a detector of electromagnetic waves. It
was also true that Braun, like Hertz, had no thought of com­
mercializing any research result that might have practical

9
applications. The prevalent attitude of most scientists was
(and in many cases remains to this day) that the results of
scientific research at the universities should be available to
all. In the case of Braun, however, this attitude would later
change when he became involved in the development of
wireless.

Solid-state Detection: the Efforts that Failed or Were

Forgotten

After the publicity following Marconi's initial successes with

wireless transmission, a number of companies in several
countries entered the new field. Germany was no exception,
where several small companies were soon competing with
each other, as well as with an already powerful British corpo­
ration under Marconi's direction. Braun was swept into this
rapidly developing situation in the late 1890's, when German
businessmen realized the need for technical support if they
were to remain competitive. He soon found himself involved
in what by modem terminology is called research and devel­
opment in private industry. Braun's interest and attitude
changed from those of the laboratory scientist to those of the
industrial inventor-developer. He experimented with wire­
less transmission and took active part in field tests of
equipment in hopes of extending its range and improving its
reliability. In fact, Braun became a major contributor to the
development of wireless telegraphy, so much so that he later
shared a Nobel Prize with Marconi for achievement in devel­
oping practical wireless communication. All this was in
addition to work on his most well known invention, the
cathode ray tube.

Braun, as well as some of the other principals in wireless de­

velopment, gradually became aware of the disadvantages

10
(and the advantage) of Marconi's original detector, an adap­
tation of the Branley coherer. It's big advantage depended
upon the initial concept of what the ideal wireless
communications system would provide. This concept envi­
sioned the automatic recording of received signals by means
of a Morse Inker or similar instrument. This eliminated the
presumed disadvantage of a human operator in the receiving
system.

The commitment to automatic recording resulted in a delay

in accepting improved detection methods, because in general
such methods were not adaptable to the automatic process.
This commitment affected the Marconi organization and the
German companies, including Braun's. Nevertheless, Braun
came to realize the significance of his earlier work and began
an effort to adapt the crystal rectifier for use as a detector in
1899. His experimental work during the 1899-1901 period
was successful, and he was able to demonstrate a working
detector, but incompatibility with the Morse Inker blocked
its application. Unlike Marconi, Braun was a major partici­
pant but not the controlling interest in his company, and he
was compelled by prevailing corporate policy to bide his
time.

Meanwhile, others had become aware of the significance of

solid-state rectification. The earliest of the investigators was
Jagadis Chunder Bose, who is known primarily for his work
in the field of plant physiology, but who also repeated the
work of Heinrich Hertz, with improvements in technique and
apparatus.

Bose was born in India in 1858, and received his higher edu­
cation at Christ College, Cambridge England, before
returning to Calcutta as a Professor of Physics at Presidency

11
College. He later founded the Bose Research Institute for the
study of plant and insect life. He was one of the first lectur­
ers to demonstrate Hertzian waves as a possible medium of
communications, both verbally and by demonstration, using
apparatus of his own design, and he suggested a number of
possibilities for the use of Hertzian waves.

Bose eventually lost interest in the subject, but not before

progressing far enough to recognize the usefulness of solid-
state rectification as a means of detection. This is evidenced
by his filing a U.S. patent application titled Detector for
Electrical Disturbances, on September 30, 1901. The patent
was not issued until March 29, 1904 as No. 755,840, but the
early filing date indicates that he was active at about the
same time that Braun was unsuccessfully trying to get ap­
proval of his own crystal detector.

The apparatus described in Bose's patent was not primarily

intended as a radio frequency detector, but it was described
as being readily adaptable for that purpose. The crystal de­
tector was described in general terms as a substance
possessing certain electrical characteristics. It was later de­
termined from other information sources that Bose had
actually used two dissimilar crystals of galena, in contact
with each other to provide rectification. Had he fully cov­
ered this part of his work in the patent, he would have
anticipated the concept of a crystal detector invention of that
type which was not patented until several years later. How­
ever, his invention did meet the requirements for a practical
device, although he did not produce one. Thus it was left to
others to take that step.

The first solid-state detector to be useful in receiving appara­

tus was patented by Sidney G. Brown. This was filed June 3,

12
1904, just a few months before the Bose patent was issued.
The detector was actually a solid-state electrolytic detector,
but its macroscopic characteristics were so similar to those of
a crystal detector that it was later catalogued with the latter
for sales purposes. The active substance was a pellet of
compressed lead peroxide powder, held between a platinum
plate and a blunt lead point.

This detector was used for a time in at least one commercial

wireless station in England, and an Electro Importing Wire­
less Catalog of 1918 was still advertising the pellets, along
with crystals for detectors, at $0.30 each. It was a true in­
vention in every sense of the word, yet it never became
popular and did not change the direction of technological de­
velopment. The most likely reason for the lack of success
was simply that it did not have the sensitivity associated with
either a liquid type of detector or the later crystal detector.

Louis Winslow Austin was another.'.scientist who investi­

gated rectification, and was the last to enter the scene before
the burst of activity that resulted in successful detectors. He
was bom in Orwell, Vermont, in 1867, and attended univer­
sities both in the United States and abroad to earn his
doctorate, become an Assistant Professor, and do research.
He became associated with the United States National Bu­
reau of Standards in 1904, beginning a long career that led to
a number of contributions to radio science.

In the early 1900's Austin discovered that some materials

which were thermoelectrically sensitive also acted as rectifi­
ers. He incorrectly believed that rectification was caused by
the thermoelectric effect, but this was understandable in view
of the limited knowledge of the time. He was not alone

13
either; a substantial number of scientists considered thermo­
electric effect as a likely explanation.

Pursuing his ideas, Austin developed a detector consisting of

a roller of silicon or aluminum, held under slight pressure
against the surface of a block of tellurium. The tellurium
available at that time was relatively impure, and its high
thermoelectric power was a beneficial effect of the impuri­
ties. Austin very likely chose tellurium because of that high
thermoelectric power, and he named his invention a thermo
detector. He later revised his opinion when experimental re­
sults became available which showed that thermal effects
were not the cause of rectification.

Austin was with the Bureau of Standards in 1906, when he

applied for two detector patents, one for silicon against tel­
lurium, the other for aluminum against the same element.
Whether he was hoping for direct involvement with com­
mercial applications, or was simply protecting his interest in
the invention is not known. The fact that both British and
United States patents were issued suggests that he attached
considerable importance to his work. However, once again
there was no further activity resulting in practical applica­
tions, at least in the United States. It should be noted
however, that in Germany a similar detector called the Bronc
Cell came into wide usage. The latter consisted of a finely
adjusted contact between tellurium and graphite.

Solid-State Detection: the Prelude to Success

Braun, Bose, Brown, and Austin were all active at about the
same time that business decisions and technology had the
effect of delaying further development. Two conditions were

14
instrumental in producing that effect and were mentioned
earlier: the desire for automatic recording, and a commitment
to the coherer type of detector. The bias in favor of automatic
recording was particularly prevalent in Europe and England,
although the United States Navy was also a proponent of that
method for a time. In the latter case it has been suggested
that such a system was believed to offer the advantages of
less training for recruits, and fewer security problems. In
any case, this favoritism naturally discouraged developments
of detectors which were not compatible with automatic re­
cording, and the crystal detector was one of these.

The coherer met the comparability requirement for applica­

tion with automatic recording. It was also an attractive
choice for another reason. There was an erroneous but
prevalent opinion that the coherer was a device of high sen­
sitivity, perhaps the most sensitive method of detection to be
found. This opinion was even employed to excuse some of
it's shortcomings, such as being difficult to adjust, unstable,
and notorious for its inability to distinguish signals from
static. In other words, it was considered to be temperamental
because of its supposedly high sensitivity, rather than its
poor characteristics.

The reason this bizarre concept was not quickly disproved,

before it delayed the progress of invention and committed
manufacturers to the wrong product, arose from the fact that
with 1900's technology there was no way of directly measur­
ing how much signal energy was available after the coherer
rectified the incoming transmission. The fraction of signal
power available at the antenna that was actually used within
the coherer served to create a low resistance path for local
battery power to drive a recorder pen or ring a bell. This
local power was measurable, but was unrelated to the

15
strength of the signal power. In fact, the only way that the
sensitivity of the coherer could be determined was by com­
parison of the performance of other detectors under like
conditions. This took much time and effort, mostly from the
cumulative experience of a number of investigators, as well
as wireless operators in the field.

A feeling that something better could be found eventually

prevailed, especially in the United States, where business
conditions encouraged a change of opinion. This occurred
because the wireless business community in the U.S. con­
sisted of several small companies struggling to survive in a
marketplace which included German manufacturers, the
American Marconi Company, and its powerful parent corpo­
ration in England. The patents held by these companies
discouraged outright copying, and there were already several
cases of litigation over claims of patent infringement by
more subtle means. On the other hand, the bigger compa­
nies, already established, were thoroughly committed to the
products based on those patents, and were understandably
reluctant to change. In short, the big companies, by their
own commitments, opened up some possibilities for the
small ones. This was a strong incentive for the invention of
new products, and for business investment to develop them.

Automatic recording was never strongly favored in the

United States, but when it was realized that the presence of
the human operator at an appropriate type of receiver pro­
duced a more sensitive receiving system, effectively
increasing the range of transmission, the advantage was ob­
vious. This was the way a small company could gain the
edge over entrenched competition, and it created an ideal
climate for American inventors. Braun and Bose could not
directly benefit from this situation, being located in other na-

16
tions, and Brown and Austin fared no better, even though
Brown had connections in both England and the United
States, and Austin was active at the Bureau of Standards. It
remained for two other men to produce the detectors which
came into widespread use.

The First Practical Crystal Detectors

Crystal detectors were not the first detectors to outperform

the coherer, and their reign of supremacy was brief. Never­
theless, for a time they were the most used means of
detecting spark transmissions, and spark was the only way
there was to transmit over great distances in the early 1900's.
Crystals were the best replacements for the coherer, and the
person most responsible for their development was Greenleaf
Whittier Pickard, who is one of the two men usually credited
with the invention.

Pickard was bom in 1877 and received his higher education

at Harvard and the Massachusetts Institute of Technology.
In 1901 he went to work for the American Wireless Tele­
graph and Telephone Company, which was the first wireless
company formed in the United States. After some months of
employment, he was working with a microphone detector
during reception of signals from one of the company's
transmitters. At this time the microphone detector was one
of the few types not covered by a patent, and thus available
to the smaller companies.

The detector was a contact-type unit consisting of carbon

against steel, in series with batteries and headphones. When
operated at full current, an audible signal was accompanied
by a background of static, commonly described as a frying
noise. While Pickard was working with this equipment, he

17
became annoyed one day at the noise. He removed two of
the three batteries in the circuit (so he thought) which elimi­
nated the noise, although signal strength was greatly reduced.
However, upon checking his circuit, he discovered that all
three batteries had been removed, and that the received signal
was due solely to the transmitter power received. Such re­
ception was contrary to all of his previous experience, and it
raised his curiosity.

Pickard's interest led him to experiment further with loose

carbon in contact with steel, then with oxidized steel in con­
tact with pointed steel. The oxide layer was fragile, which
led him to consider using iron oxide in bulk form. In Octo­
ber 1902, he tried a sample of magnetite (Fe203) and
received signals with better volume than any of his previous
arrangements. In short, Pickard was now working with a
crystal detector, discovered and tested independently of the
previous investigators. During the next three years, he tested
a large variety of specimens, including chalcopyrites, iron
pyrites, and galena, all of which were later used in successful
receivers.

By that time there was evidently a general awareness of the

possibility of mineral applications as solid-state detectors.
Pickard's notes show that he was acquainted with the theory
that detection was a thermoelectric effect, and this influenced
his choice of minerals for testing. Nevertheless, at that time
and even in later years, he was still unaware of Braun's de­
velopment of a practical detector.

Pickard continued his search for better materials, and in 1906

he was able to procure some fused silicon, which was at that
time a new product of the electric furnace. Silicon specimens
worked so well, using a point contact connection, that he was

18
ready to patent his discovery and manufacture crystal detec­
tors as a commercial product.

At about this time the second successful inventor entered the

detector development field. Henry Harrison Chase Dun-
woody, bom 1842, had graduated from the West Point
Military Academy in 1886. During his army career he was
promoted to Brigadier General, with most of his service in
the Signal Corps. His resulting experience in Army com­
munications may have been a factor in his joining the de
Forest Wireless Telegraph Company as a Vice President, af­
ter retiring from military service.

In 1906, Dunwoody was actively involved in developing a

crystal detector, perhaps upon a suggestion by de Forest, and
on March 23, 1906 he applied for a patent. This came at a
fortunate moment for the de Forest interests, because an
electrolytic detector that de Forest was using in his commer­
cial receivers was the subject of a patent infringement
lawsuit. The litigation had been in progress for some three
years, and the court finally ruled against de Forest in 1906.
De Forest desperately needed another type of detector when
Dunwoody made his discovery.

Dunwoody took advantage of the characteristics of carbo­

rundum (silicon carbide, SiC) crystals, another product of the
electric furnace. A typical batch of carborundum ranges all
the way from highly colored to dull gray specimens, depend­
ing upon their location in the furnace, and he found the less
colorful crystals to be the most promising. Dunwoody
picked dull gray as preferable. Others have referred to
greenish gray as the best choice. Dunwoody also found an­
other property of carborundum to be helpful. Elongated
crystals are not uncommon and he chose these for his

19
specimens because electrical connections could be easily
made at each end.

His method of making connections consisted of winding a

few turns of wire around the ends. Current rectification oc­
curred when there was a difference in the semiconductor
properties from end to end, which occurred in a small num­
ber of the specimens. It is evident from this that Dunwoody
was unaware of the advantage of using a point contact at one
connection. Therefore, he either was unacquainted with or
did not understand earlier results in this regard. His own re­
sults with appropriate specimens were still impressive, so he
filed his patent application on March 2, 1906, preceding
Pickard's application by about five months.

It is noteworthy that Austin filed for his patent on October

27, 1906, about two months after Pickard's filing date of Au­
gust 30, and that is not all. Braun, returning to the scene
after several years, filed for a German patent that same year.
What caused this sudden burst of activity, from which
Pickard and Dunwoody were to emerge with the benefits?

First, there was the growing awareness of the need for better
detectors and it appears that by 1906 there was at least some
general speculation that minerals could be used in this re­
spect, as witness the interest in thermoelectric theory as well
as other possibilities in explaining the phenomenon of solid-
state rectification. Could this have encouraged Austin's pat­
ent application at that particular moment? The answer is not
known at this time.

Pickard's situation is more evident. By all accounts he was at

the stage where it made good sense to obtain patent protec­
tion for a device which he intended to produce. In fact, his

20
attorney is said to have urged him to generalize his applica­
tion to cover all substances with useful sensitivity. In
actuality, the applications went in one at a time over a period
of months, each covering a specific case. As for Dunwoody,
he and de Forest needed a detector immediately, or they
would have been forced to halt receiver production after
losing the patent lawsuit against them concerning their elec­
trolytic detector. There is no doubt that they filed to patent
the carborundum detector as soon as possible.

Ferdinand Braun was also directly motivated, even though he

was an ocean away from American developments. He was
now associated with the Telefunken organization, the latter
being the result of merging his company and others by direc­
tive of the German government. The object of the
consolidation was to improve the German competitive
position relative to foreign competition, especially the British
and American Marconi.

Telefunken management realized that better equipment was

necessary if they were to remain competitive. In the case of
receivers, it was decided to develop Braun's crystal detector,
even though it meant the end of previous commitments.
Two other considerations may have contributed to this deci­
sion, even though they were based on rumors later found
untrue. The first of these was concern about another German
invention.

A young researcher named Robert von Lieben was working

upon a three element tube, predating de Forest's invention of
the triode. Von Lieben's procedure and goal were somewhat
different, because he was interested in amplification rather
than detection. Nevertheless, Telefunken management per­
ceived von Lieben's work as a possible threat against their

21
current detector technology, and they wanted to counter it.
(Telefunken and the von Lieben interests later cooperated in
development of the tube.)

A second concern stemmed from ongoing press commentary

about the work of Nikola Tesla in the United States. Tesla
was somewhat later relegated to the rank of visionary, but at
that time it was still fresh in mind that he had single-hand­
edly revolutionized the power industry by inventing the
polyphase AC generation and distribution system. Tesla was
never shy about promoting his ideas and had conducted
spectacular demonstrations involving high frequency power
and wireless transmission.

As early as 1902, Tesla had directed construction of a huge

tower near the town of Shoreham, Long Island, and he had
announced a forthcoming system, a "World System" of mul­
tiple uses of wireless power transmission, including
transoceanic broadcasting. Tesla was encountering serious
financial difficulties by 1906, but until just a few months
earlier there had been an almost continuous stream of favor­
able publicity, marred only occasionally by bizarre
speculation. It has been suggested that Tesla's concept of a
World System, which included a "World Wireless System"
as part of the package, was also a factor in Telefiinken's de­
cisions. Tesla did, after all, show a major construction
investment on Long Island, and there were eyewitness
reports of tests in progress that demanded very large amounts
of power.

Whatever the reasons, Telefunken management gave Braun

the authority to proceed, and he chose to patent a detector
using psilomelane, a fairly common mineral composed
mainly of manganese dioxide. It is not clear whether his

22
choice was based upon locally available mineral deposits
with exceptionally good properties, or whether his specimens
of other minerals were of poor quality. In any event, the
mineral he selected never came into use in the United States.
It's post-patent history in Germany is not known at present.
It is known that Telefunken later used carborundum crystals
for at least some of its detectors, after reaching cross licens­
ing arrangements with the patent holders.

The applications for patents and immediate practical usage

made 1906 the year that the crystal detector arrived as a
commercial receiver component, and this, of course, made
the crystal set a reality. The latter soon became the most
important type of receiver in use, and it remained so for sev­
eral years. This was true in spite of the fact that the Marconi
Company in England was already using a vacuum tube (the
Fleming Diode Valve) and that de Forest had already applied
for a patent on a similar design, resulting in more litigation.

Further, it was as early as January 28, 1907 that de Forest

applied for his triode Audion patent. It was only later rec­
ognized for what it really was, the forerunner of the high
vacuum triode that radically changed the entire communica­
tions industry. Curiously, even de Forest did not seem aware
of its importance for application as a detector until about
1910-1911, and so the crystal detector dominated the scene
until World War I. It was still used throughout that conflict,
concurrently with other alternatives. It was not unusual to
see a manufactured receiver complete with one or more
crystal detectors, plus a panel mounted vacuum tube and
suitable provisions to switch from one to another.

23
Continuing Development

The first patents did not signal an end to crystal detector de­
velopment. Pickard continued to investigate methods and
materials for improvement. He had chosen the combination
of a pointed brass wire pressing against silicon, which he
called a Perikon detector (Perfect Pickard Contact). Later the
term came to be more closely associated with another
Pickard invention: a blunt pointed crystal, held in contact
with the surface of a second crystal of different composition.
He procured a registered trade name for the term, but the
name is now generally used for any combination of two
crystals in a Perikon type arrangement, regardless of origin.
Pickard patented the Perikon combination of zincite and
chalcopyrites, which was used to some degree along with the
other detectors.

The de Forest Company also continued work on the crystal

detector. The first method of making connections used by
Dunwoody was soon replaced by a more practical design by
an employee of de Forest, C. D. Babcock. The new device
was a holder which clamped a crystal fragment between two
flat metal surfaces with a provision for changing the clamp­
ing pressure by means of an adjustment screw, This
arrangement also permitted changing contact locations by
loosening the screw and moving the crystal. Babcock ap­
plied for a patent on August 27, 1906, only five months after
Dunwoody's application date, so it is evident that the de
Forest interests were continuing their development effort.

It is also evident that de Forest was still in trouble.

Babcock's design provided the equivalent of a point contact
by relying upon the jagged surface of the crystal fragment
touching the metal surface of the holder. A problem arose

24
from the fact that this was true of both connections rather
than just one, and for mechanical stability as many as three
points at each connection would usually exist. In short, the
de Forest investigators were still unaware of the best
mounting method for a crystal: a point contact at one con­
nection and a wide area surface contact at the other.

The magnitude of the problem became all too obvious when

the de Forest Company attempted to produce the carborun­
dum detector for commercial applications. Only about one
in ten of the production units performed satisfactorily, and
nobody understood why. The solution was quite appropriate
to the occasion: Pickard was retained as a consultant to solve
the dilemma.

There is some question as to just when Pickard himself had

recognized the importance of a wide area electrical contact at
the base of the crystal and a point contact at the other, but
there is no doubt about his success with the carborundum
detector. His new design consisted of a stiff pointed contact,
held under pressure against the active surface of a carborun­
dum crystal mounted in a button of solder. This provided the
optimum contact condition at both connections. With these
changes, the carborundum detector was produced with a
much lower rejection rate. He also introduced the bias bat­
tery as a method of considerably improving detector
sensitivity, and he later received a patent on his innovations.
This improved detector was used for several years by the de
Forest Company, and by its successor, The United Wireless
Telegraph Co.

Pickard and his attorney, Phillip Farnsworth, started a small

company to manufacture and sell crystal detectors. In Feb­
ruary of 1908, they, together with a retired army colonel

25
named John Firth, incorporated the company as the Wireless
Specialty Apparatus Company. They produced crystal detec­
tors and other components at first, and later expanded to
manufacturing complete radio receivers. Their detectors
were tailored for marine applications, which at first consti­
tuted their principle sales market, although the U.S. Army
Signal Corps was also a customer. Their equipment was of
the highest quality and is much sought after by present day
collectors.

There was another outlet for sales which began to assume

some importance during the first decade of this century. A
growing number of amateurs in the United States were ex­
perimenting with wireless. Most of these were boys and
young men with limited funds, and who often built their
equipment from whatever was at hand. One of the few
things they could afford to buy outright was a good crystal,
which was an inexpensive as well as sensitive detector. As
they matured and had more money, they began buying detec­
tor stands and other manufactured components, This
amateur market appears to have developed very early after
the introduction of the crystal detector in late 1906. As early
as 1907, one of the pioneers of radio, Lloyd Espenshied,
himself an early amateur, was already using a crystal detec-
tor.

During this period, at least one, and very likely a number of

experimenters, discovered that using a fine wire catwhisker,
rather than a heavier pointed wire, improved the sensitivity
of a number of different detector materials. In spite of that
improvement, nobody filed a patent application for the cat-
whisker contact until Pickard took advantage of the situation
and filed in 1911. Years later, Pickard made the following
comments concerning the invention of the catwhisker: "I

26
have no idea who first used a fine wire contact with materials
requiring low pressure contact, and by 1911 my company
had produced and sold a number of fine wire contact galena
detectors to the U.S. Navy, and in the same year I filed a pat­
ent application for the 'catwhisker' detector which issued
later as U.S. Patent 1,104,073. I have no doubt that some­
where around 1911, or perhaps even earlier, the same idea
occurred independently to others." The catwhisker had one
big disadvantage. Anyone who has wrestled with the ad­
justment of a catwhisker, first searching for a sensitive spot
on the crystal, then maintaining good contact, will
understand why the catwhisker was never very popular on­
board ship. It is very sensitive to vibration and shock and
easily knocked out of adjustment. Marine operators faced a
formidable problem working with the device, especially in
rough seas.

Some of the minerals used for detectors were subject to still

another drawback, which the use of a catwhisker tended to
magnify. Heavy static or operation near a transmitter some­
times destroyed the sensitivity of certain materials at the
point of contact. Some onshore stations were built with the
receivers located at a distance from the transmitter, at least in
part because of this problem, but this could not be done on a
shipboard installation. Of course, these difficulties did not
prevent the growing demand in the amateur sales market, and
for other limited applications.

The Competitive Years, and the Role of the Crystal

Detector

The small American Companies, which struggled through

the first few years after the turn of the century, either failed,
changed hands, or consolidated, so that the business scene

27
had radically changed by 1910. One of these changes was
the acquisition of the de Forest Company by a new company,
The United Wireless Telegraph Company, followed by rapid
growth of the latter. Most of the de Forest assets went to
United Wireless, including Dunwoody's carborundum detec-
tor. In the brief period between 1907 and 1911, United
Wireless grew to the point where it was the dominant
company in the United States, having far outstripped it's
principal rival, the American Marconi Company, a subsidiary
of British Marconi.

The Marconi interests were producing receiving equipment

which depended upon the Fleming Valve, a vacuum tube di­
ode which performed better than their previous detectors.
Unfortunately for Marconi, it had become evident that the
Fleming diode was not as satisfactory as the crystal detector,
and the same problem was also evident in comparing much
of the other Marconi equipment with the competition. How­
ever, the sheer size of British Marconi and its subsidiaries
meant that the Fleming Valve was more important world­
wide than the crystal detector. All of this was soon to
change, because of an important event that took place in
1912.

At that time, American Marconi filed suit against United

Wireless for using certain Marconi patents without permis­
sion or licensing agreements. There was considerable
justification for doing this, and United Wireless was well
aware that they would lose the resulting court decision.
Therefore, they did the only sensible thing they could do un­
der the circumstances, which was to plead no contest and try
to reach an out of court agreement with American Marconi.
An agreement did result, in which American Marconi ac­
quired virtually all of the assets of its much larger rival, and

28
suddenly had the advantage of a near monopoly in the United
States for the manufacture and sale of wireless equipment.

One consequence of this dramatic turn of events was that the

carborundum detector became available not only to Ameri­
can Marconi but to the even larger parent corporation in
England. In fact, the Fleming Valve was soon replaced by
the carborundum detector, and the latter became the standard
detector throughout the Marconi business empire for several
years.

The Wireless Specialties Company was a small independent

corporation not directly affected by this situation, and it was
as its name stated, strictly a specialty company. It continued
to supply the crystal detector types protected by Pickard pat­
ents, and to build and sell high quality receiving equipment.
There was no effort to penetrate the high power transmitter
segment of the wireless business. In those days, manufacture
and construction of large transmitter installations was analo­
gous to the "heavy industry" portion of business in general,
due to the huge size, power requirements, and complexity of
supporting facilities.

American Marconi continued using the carborundum detec­

tor until 1913, at which time it began offering a receiver of
improved design employing both a carborundum and a
cerussite (PBCO) detector as alternative choices. The
cerussite crystal was claimed to be as sensitive as a galena
detector, which is now considered to be the most sensitive
available substance. Most of the well known detector
materials were in use by 1910-1913, although it would be
many years before germanium and high purity, precisely
doped silicon would be available. Pickard's and Dunwoody's
inventions both survived and were commercially successful.

29
The contributions of Braun and Austin were never utilized in
the United States, and Bose's work had not affected the
course of events. The understandable but misleading tale
that Pickard and Dunwoody were the sole investigators in the
field probably stems from that situation.

The major part played by the crystal detector just before

World War I resulted from several conditions. First, the
Fleming Valve was not as sensitive as a good crystal. This
was one reason why the Marconi companies took the oppor­
tunity to use carborundum after the acquisition of the United
Wireless assets. Further, the crystal detector was easy to
produce, much less fragile than a glass vacuum tube, and did
not require considerable power to heat a filament. The de
Forest Audion, first as a diode similar to the Fleming Valve,
and shortly thereafter as a triode, had been invented almost
concurrently with the carborundum detector, but the triode
version had serious problems. It was not, in fact, recognized
at first as a practical device. This was in spite of consider­
able publicity and promotional efforts. De Forest himself
delayed several years before proceeding further with it, and
even then found it to be unreliable. Audions were never
mass-produced with closely similar characteristics, and the
characteristics of any one tube sometimes changed unexpect­
edly during use. Competition between the crystal detector
and the triode was not important until several companies in
the United States and abroad realized the possibility of de­
veloping a high vacuum triode for an amplifier tube as well
as for simple detection. This did not occur until about 1912.

Other devices had been proposed as detectors, and some

were patented and used in varying degrees. Marconi's mag­
netic detector was reliable, but it became obsolete because it
was too insensitive. The electrolytic detector was successful

30

i
 using liquid electrolyte, but the latter, usually nitric acid,
spilled easily with unpleasant results. The solid electrolyte
version did not have the spillage problem, but it remained in
obscurity for unknown reasons. It can be speculated that it
was too insensitive, as was mentioned earlier. The hetero­
dyne detector, proposed in 1902, showed great promise, but
its application was delayed during the 1902-1910 period
because of lack of satisfactory operating components. An
electromechanical system called the tikker (or ticker) did
reach a degree of limited use. A number of other devices
were patented, but they were found to be impractical.

World War I brought with it a transition period with respect

to the crystal set and other receivers. An unusual state of
coexistence was evident among those in use at the outset of
the war. These remained in use concurrently, and in the case
of crystals and tubes, both were used in new receiver designs
as the war progressed. Improvements in vacuum tube de­
sign, and the invention of circuitry which took place just
before and during the course of the conflict, later changed
this state of affairs. Specifically, the invention of the regen­
erative detector, actually a receiver circuit using the
improved high vacuum triode, was very likely the principle
reason for this change. Such receivers showed dramatic im­
provement in performance compared with the crystal set.

By the end of World War I, the superiority of the regenera­

tive receiver, and also the advantages of amplification using
the new tubes, were widely recognized. This new technol­
ogy was hampered at first by the legal restrictions of patent
law and a storm of litigation, but this did not prevent the ob­
solescence of the crystal set. Its days were numbered as far
as the "traditional" business of point-to-point communica­
tions was concerned. In that particular field, especially for

31
l
shipboard installations, the crystal set assumed the function
of an emergency back-up unit, available when all else failed.
This was a minor application, but a small demand did exist
for many years. It was a post-World War I phenomenon of a
different nature that extended the life of the crystal set as an
important commodity.

A New Industry

The amateur market grew rapidly after the war, providing a

sales outlet for crystals, but what really rescued the crystal
set from obscurity for a few more years was broadcast radio.
1920 marked the beginning of the growth phase for this new
industry and of its importance as part of the American scene.
Within just one decade, broadcast radio created an enormous
market for home receivers and affected the habits and
customs of the general population.

Crystal sets were very popular at first, because they were

inexpensive and easy to operate compared with more elabo­
rate radios. For example, in 1922 manufactured crystal sets
were available for $25 and up, whereas the least expensive
one tube receiver sold for about $75. Crystals were also used
in some sets together with tubes. Most such usage occurred
during a brief period in the early 1920's when interest in
radio assumed some of the characteristics of a fad or craze.
At times the demand for radios outstripped the supply.
Unlike a fad, however, home radio sales matured into a
permanent component of the total consumer market. This
created a major manufacturing and sales industry, which
supplied a public that became more and more sophisticated
in its preferences as time went by. This became an important
factor which affected the evolution of the radio.

32
Most broadcast band receivers built before 1923 were con­
structed at home, and ranged from simple crystal sets to one
to three tube designs. Inexperienced family members often
found them difficult to operate, and the sets using tubes often
required bulky and expensive batteries. Manufactured radios
were a small fraction of the market at first, in spite of the
great demand, because it took time to organize a new com­
pany, and then design, test, and set up a production line.
Many new companies were started, but it was not until 1925
that annual commercial production equaled the number of
home built radios during the year.

Consumer demand favored the latest improvements in con­

venience and performance, and this, plus rising competition,
pushed a rapid development of better receivers. That trend
doomed the crystal set as a major market seller. The crystal
set was acceptable while there was just one station in town,
but as more and more stations came on the air, as multiple
frequencies were assigned, and when a fad for long distance
reception blossomed, the inferiority of the crystal set became
obvious. It was too insensitive, selectivity was poor, and
listening with earphones was awkward for large families.
The multi-tube radio was the big winner here, even though it
meant the end of the era when most receivers were home
assembled.

After 1921, the manufactured radio dominated the retail sales

market, and its design and production had become an impor­
tant engineering specialty which continued to evolve during
the 1925-1930 period. The biggest single improvement was
the introduction of AC operated radios which performed
well, and at the same time could be purchased at a reasonable
price by the average family. That milestone was reached in
1928, and by that time the crystal set was a thing of the past,

33
as was the crystal detector itself in most cases. In fact, by
1930, products using crystal rectifiers for any purpose were a
very minor percentage of the total market for all types of
radio equipment. A minor market did exist, nevertheless,
and deserves discussion. Beginners in radio, as well as
hobbyists, created a small but stable sales outlet for many
years. There was also activity in the sales of a miniature type
of radio best described as a toy or novelty. These topics are
covered in the next Section.

Survival as a Specialty

Crystal set circuitry was an attractive method to introduce

students to electronics for many years, and the diode detector
circuit is still in the curriculum. The detector stand, crystal,
and catwhisker have long since been replaced by a germa­
nium diode, but the basics are still the same. What is
important to the suppliers is that the market may be small,
but at least it is stable. Barring utter catastrophe, each year
sees a new class of beginners, even though the numbers may
fluctuate with demographic trends. In this situation a solid-
state diode and associated circuitry continues to play a very
minor but very long duration teaching role for now and the
future.

Another small but more dynamic market has been the supply
of crystal set related hardware and information to hobbyists
and collectors. It is reasonable to relate the origin of such
business to activity of the early 1920's, when the crystal set
was popular as a home radio. Many were purchased part by
part and assembled at home, so it was profitable for suppliers
to stock appropriate materials and components. As radios
were improved, the home-brew crystal set market shrank, but
there were still numerous sources for parts through the

34
1930's, and even after World War II. The local radio repair
shops had new stock, and there was an abundance of used
and discarded radios, available at nominal cost, which could
be sold for parts.

There was another important source which has since disap­

peared, along with the local radio shop which is now all
television. This source consisted of a number of national and
regional mail order suppliers, and for over four decades they
offered just about anything that might be needed. The na­
tional suppliers included such names as Allied Radio
Corporation, Bemstein-Appleby, Concord, Lafayette, and
Radio Shack. Most of the larger suppliers remained active
until the 1970's or '80's, but they eventually went out of busi­
ness or changed their product lines to concentrate on
complete receivers, high fidelity equipment, industrial
electronics, and even non-electronic appliances and
equipment.

The demise of the mail order components business was not

due to the decline of the crystal set, which had occurred
many years earlier. Even before 1930, most parts sales were
to do-it-yourselfers working on more elaborate equipment.
There was even a modest business in television parts and kits
for a time after World War II. There has been some specula­
tion concerning what caused the decline in parts sales. One
hypothesis suggests that the rise of solid-state technology
resulted in such compact, highly efficient, and professional
looking products, that the home brew enthusiasts felt it im­
possible to match or exceed the commercial equipment. At
present there is an intense interest in electronics, but often
the interest lies in assembling a complete system, using
building blocks such as tuners, amplifiers, playback units,
and so forth, with no curiosity as to what goes on within the

35
individual blocks. Therefore, "black boxes" still sell, but not
the parts inside.

Whatever the reason, there was a period of years during

which crystal set components were scarce. Hobbyists should
thank the growing collector interest in early day radios for
improving the parts situation, even while they are removing
older radios from availability. Collecting, which includes
restoration, reproduction, and replicating, became popular
enough during the 1980 decade to encourage small busi­
nesses to begin supplying parts and information again. At
first, supplies were from "new-old-stock" or were used.
However, now in the 1990's, there is a limited production of
components and small hardware needed for crystal sets.

One small mail order business specializing in the crystal set

field began during the unlikely depression period in about
1932, and persisted through the lean years that followed.
This was Modem Radio Laboratories (MRL) now located in
Minnesota, although originally a California based company.
MRL was founded by Elmer G. Osterhoudt, a 1920's era
wireless operator, amateur, and later a radio serviceman.

After establishing MRL as virtually a one man operation,

Osterhoudt became known as an expert in practical knowl­
edge of the crystal set field. He published and copyrighted
much literature on the subject, all tailored for both novice
and experimenter, which was sold through his company.
This included pamphlets, a magazine-type periodical, cata­
logs, and other information. The catalogs offered a wide
range of components to his customers, which were ideal for
crystal set applications. Also sold were kits of his own
design.

36
MRL survived the years when larger companies left the mail
order business, and even produced its own simplified detec­
tor stands and switch point levers when these became
otherwise unavailable. Selected components, which could be
substituted for hard-to-find parts needed for new construc­
tion, were also located and featured. Osterhoudt's death
following a 1987 automobile accident caused the company to
suspend its operations for a time, but, about a year later it
was sold and moved to Minneapolis to resume operations.

Another, rather different type of business lingered on for

many years after a strong beginning in the 1920's. Small
crystal sets were cheap and sold well, and some of these were
reduced to the minimum possible circuitry. The dominant
company in the field was the Philmore Manufacturing Com­
pany, which had started earlier as the Ajax Products
Company before its incorporation in 1925. Their cheaper
products were small open type sets, with outside-mounted
detector stands or holders for the crystal. These sets were of
a size and design that actually placed them close to a novelty
and toy category, and later their entire product line veered
toward that type. They did continue to sell separate compo­
nents, chiefly a detector stand, which was still available as
late as 1956.

There were other companies in the field at one time or an­

other during the long span of Philmore's activity, some of
them actually doing business before Philmore started. None
seems to have lasted as long, however. There was also a true
novelty and toy market in the 20's, supplied by several small
businesses. Another wholesale outlet for the manufacturers
was the supply of premiums advertised to boys as an in­
ducement to sell a product or eat a certain brand of cereal.

37
Good quality crystal sets disappeared from the stores as the
public demanded better performance, but the novelty and toy
market continued for over thirty years. In fact, it may not
have reached its peak until after World War II, when adver­
tisements promoted tiny crystal sets as a "new technology."
Except for that emphasis, the type of advertisement remained
almost the same during the entire era of novelty manufactur­
ing. Even in the early 20's, advertisements picture a boy
grinning with excitement as he listens to his tiny set, also
prominently pictured, and the same general format was used
for years. Early promotions also emphasized some of the
set's technical details, but later it was simply stressed that no
batteries were ever needed. The low price was also featured,
and for premiums the "free" inducement appeared in big bold
letters. The fact that an external antenna and ground were
required was either not mentioned at all or downplayed.

The Philmore company was in the novelty and toy market

from the early days, but a later rival appears to have out pro­
duced it in more recent years. Western Manufacturing
Company of Kearney, Nebraska made these radios using a
number of brand names and small company names for firms
that they either owned outright or controlled. Some other
companies also participated, although they appear to have
been less prolific.

There were manufacturers that took advantage of what was

probably the most famous miniature radio of all time, the
Dick Tracy comic strip wristwatch radio. The latter was a
fictional two way communications set, but Da-Myco Prod­
ucts Company of New York City made a wristwatch crystal
set and advertised it as the Dick Tracy Wrist Radio. Da-
Myco advertised the receiver as a toy, but there were some
competitors who were not quite so truthful.

38
More generally speaking, the appearance of most novelty and
toy radios evolved from the early day arrangement that usu­
ally had an exposed crystal detector stand or holder, to a
modem looking little box, often made of colored plastic, and
shaped either like a miniature table model radio, or to some
novelty form. This evolution became practical for even the
cheapest sets with the advent of the solid-state diode, de­
scribed in the next Section.

Mass-Produced Rectifiers

The decline of the crystal detector in its original form did not
mean the end of the crystal rectifier. Two important devel­
opments in electronics saw to that. First came laboratory
research in the 1930's which led to radar for military appli­
cations. During that time it became evident that the crystal
rectifier would be the only satisfactory solution for certain
applications at microwave frequencies.

After initial work in the laboratory with hand-made rectifiers,

a concerted effort was made to develop a miniature fixed
diode package that could be mass-produced for operational
radar sets. The first such diodes used silicon. Later on,
diodes were developed using germanium, the semiconductor
element not available in the early days of radio. The engi­
neering required to fabricate packaged diodes was a
considerable achievement in itself, and the research and de­
velopment effort to produce silicon and germanium with the
required properties was even more impressive.

The minerals and other substances for early day crystals were
used in their natural state, or as produced, and the best per­
forming crystals were chosen for service. Selection was

39
often made by the radio operator, searching for sensitive
spots on each crystal. Advertisements for mounted crystals
sometimes specified them to be pretested, but this could
mean about anything in those days. Either method would be
impractical with modem diodes.

The substances for modem applications had to be highly pu­

rified and then doped with very small but precise amounts of
one or more other elements, in order to meet rigid quality
control standards. Mass-production without such standards
would result in unacceptably high rejection rates. The time
and expense needed to develop acceptable silicon and ger­
manium crystals probably exceeded all of the previous
efforts that resulted in the original crystal detectors.

By the time that World War II ended in 1945, the production

capability was at hand to provide crystal diodes for civilian
applications. This meant that the germanium diode was con­
veniently available when the second major development in
electronics that would affect its use took place. This, of
course, was the invention of the transistor, by William
Shockley, John Bardeen, and Walter H. Brattain, all at the
Bell Laboratories subsidiary of American Telephone and
Telegraph. The invention was first demonstrated on Decem­
ber 23, 1947, and it was considered so significant that the
trio was jointly rewarded the Nobel Prize for Physics in
1956.

Like so many inventions, the transistor did not immediately

revolutionize an industry, electronics in this case. It took
lengthy efforts to improve the transistor and to develop
mass-production methods. It was late 1951 before the
manufacturing subsidiary of AT&T began full-scale
production, joined by the Raytheon Corporation a little later.

40
■—

In fact, it was several years before the transistor became a

major competitor with the tube. During that interim period
in the 1950 decade, sales of crystal diodes developed to some
degree on their own for various requirements.

The first transistorized radios were portables, where the ad­

vantage over tubes was highly evident. These were first
produced in the 1950's, and were not the sudden and over­
whelming success that some had envisioned. The general
public was not much interested in portables at that time, so
only modest sales were realized. The trend was there, never­
theless, and within a few more years fully transistorized
radios and television sets were the rule rather than the ex­
ception. This was the great opportunity for widespread use
of the germanium diode for detector as well as other
applications, since it made no sense to retain the vacuum
tube for that one purpose. The silicon diode also made a re­
appearance, for uses where its properties were more suitable
than those of germanium.

Since the 1950's, literally hundreds of different solid-state

diodes have been developed for as many specialty applica­
tions, and diode detector usage continues. The even more
recent development of integrated circuitry has added another
physical form and still smaller size to the array of diode con­
figurations, but the same basic principle still applies. In the
long run, it has turned out that the nineteenth century discov­
ery of unilateral conductivity did not remain the scientific
curiosity that became the fate of some other discoveries of
the time, but led to the broad applications of the present day.

41
BIBLIOGRAPHY

Aitken, Hugh G. J., Syntonv and Spark, the Origins of Radio.

Princeton University Press, Princeton, New Jersey, 1985.

Bamouw, Erik, A Tower in Babel. Oxford University Press,

New York, 1966.

Blake, George G., History of Radio Telegraphy and

Telephony. Chapman and Hall, London, 1928.

Briggs, Thomas H., The Triode that Pre-dated de Forest:

Robert von Lieben and the LRS Relay. The AWA Review,
Vol. 5, 1990, p. 45.

Cheny, Margaret, Tesla. Man out of Time. Dorset Press,

New York, 1989.

Constable, Anthony, Early Wireless. Midas Books, 12 Dene

Way, Speldhurst, Tunbridge Wells, Kent, England, 1980.

The Continuous Wave. Technology and American Radio.

1900 - 1932. Princeton University Press, Princeton, New
Jersey, 1985.

Denk, William E., Jagadis Chunder Bose and his Galena

Detector. The Old Timer's Bulletin, Vol. 27, No. 1, June,
1986, p. 16.

Douglas, Alan, The Crystal Detector. IEEE Spectrum, April,

1981, p. 64.

Douglas, Susan J., Inventing American Broadcasting. 1899 -

1922. The Johns Hopkins University Press, Baltimore, 1987.

42
Dunlap, Orrin E., Jr., Radio's 100 Men of Science. Harper
and Brothers, 1944. Reprinted by Books for Libraries Press,
Freeport, New York, 1970.

Espenshied, Lloyd, 1907 Amateur Station. The Old Timer's

Bulletin, Vol. 31, No. 1, May, 1990, p. 11.

Kurylo, Friedrich, and Charles Susskind, Ferdinand Braun.

The MIT Press, Cambridge, Mass., 1981.

Lewis, Thomas S. W., Empire of the Air. Harper Collins

Publishers, New York, 1991.

Maclaurin, W. Rupert, Invention and Innovation in the Radio

Industry. MacMillan Co., New York, 1949.

Mayes, Thom L., Wireless Communication in the United

States. The New England Wireless and Steam Museum, Inc.,
East Greenwich, Rhode Island, 1989.

McNicol, Donald, Radio's Conquest of Space. Murray Hill

Books, Inc., 1946. Reprinted by Amo Press, New York,
1974.

Pickard, Greenleaf Whittier, How I Invented the Crystal

Detector. Radio Age, Vol. 2, No. 10, Dec. 1976, p. 1.
Reprinted from Electrical Experimenter, Aug., 1919.

Schiffer, Michael Brian, The Portable Radio in American

Life. The University of Arizona Press, Tucson, Arizona,
1991.

Sievers, Maurice L., Crystal Clear. The Vestal Press, Ltd.,

Vestal, New York, 1991.

43
Thackeray, Desmond P. C., When Tubes Beat Crystals:
Early Radio Detectors. IEEE Spectrum, March, 1983, p. 64.

44
 V CHAPTER 2
FUNDAMENTALS OF RECEIVER
SYSTEMS

Some Comments on Definitions

The crystal set certainly played its part in

f^i Q q) the history of radio, as told in chapter 1.
o
to
Yet it lives on in each and every receiver
and communication system that exists today. Crystal detec­
tors and the crystal set as a whole laid the foundation on
which all receiver systems are based. Hence, in order to de­
sign these systems, or to design an advanced crystal set
today, these fundamentals are still the first principles that
you must understand. That is how this chapter fits in; it is
the transition between the idea of the crystal set (and its his­
tory) and designing and/or building sets. This is the theory
class. The lab, or design chapter, follows.

The terms "crystal set" and "crystal detector" originated at

the time when detector materials were often naturally occur­
ring crystalline substances. Detectors were not always
natural substances however; amorphous, semimetallic, and
fused solids were sometimes used. The term "mineral detec­
tor" has been employed, but the same exceptions still apply.
In short, crystal, mineral, and other words that have been
suggested are all misnomers. Nevertheless, the word crystal
is so firmly attached to the subject that it will continue to be
used here in spite of its shortcomings.

There has been some lack of agreement in defining what a

crystal set actually is. It is defined here as a receiver em­
ploying a crystal detector, which operates only on signal
power delivered by the antenna, the only exception being for

45
those detectors which require bias voltage at the crystal to
improve their sensitivity. The exception dates back to the
earliest days of the crystal detector, when it was found that
bias voltage on a carborundum detector improved its per­
formance.

There are reasons for the narrow limits on the definition of a

crystal set. One of the most obvious comes from the fact that
no one thinks of a transistor radio using a solid state diode
detector as a crystal set, so it is not considered as such. But,
consider a less clear-cut example. Is a crystal set still a
crystal set when an audio amplifier is connected at its output
terminals? As defined here, the answer is still no. Amplifi­
cation with the help of external power would improve the
performance of almost any crystal set, but for some the ex­
citement of distant reception by signal power alone would be
missing. Further, if improved performance is the goal and
amplification is acceptable, why stop at adding audio to an
existing crystal set? Why not optimize the detector-amplifier
combination as a system? The design challenge is there; the
resulting improvement would be quite evident. It is simply a
different category of challenge.

On the other hand, suppose that an active element such as a

transistor is present, but it is not externally powered. Here,
this is still a crystal set. The system is limited to power
supplied by the antenna; the achievement of distant reception
without external power is still possible. Although no one has
demonstrated that an active element can be used to advantage
this way, there are some possibilities that can be investi­
gated.

A self-powered transistor circuit with crystal detection has

been suggested. One such arrangement consists of two

46
crystal sets in one: one tuned to a local station to develop a
supply voltage and the other as the crystal set that includes a
transistor audio amplifier. The voltage obtained by rectifica­
tion of the local station's signal is used to "power" the
transistor of the crystal set. The sets would share the same
antenna. Still other possibilities are the use of an active ele­
ment as an impedance matching device, or as a means of
improved coupling between a tuned circuit and the detector.
These are all crystal sets as long as the ban on external power
is observed.

Concerning other definitions used here, most of the termi­

nology of modem electronics applies, and the reader with a
background in basic electronics will usually feel at home.
There are a few instances, however, where component names
and units of measurement have changed over the years.
Also, some components and circuits are used in crystal sets
which are rarely if ever found in present day equipment.
These components will be defined as needed.

Signal Reception Requirements

No specialized knowledge is necessary to understand that a

device which is completely dependent on transmitted power
needs to capture as much of that power as possible for best
operation. In general, this need translates into operating with
a long antenna and a good ground connection. These are
requirements that differ dramatically from what is satisfac­
tory for today's typical receiver. The antemia and ground are
so vital to the operation of a crystal set that they are treated
here as parts of the overall system.

The basic crystal set antenna is the inverted L, consisting of a

long, straight, horizontal length of wire which picks up the

47
transmitted radio frequency signal, and a short vertical
length, which is the lead-in. In early day terminology, the
horizontal wire was called the aerial, and the aerial plus lead-
in combination was called the antenna. Modem wording is
used here.

Antennas

The inverted L antenna is somewhat directional, being most

sensitive to signals arriving parallel to and from the distant
end of the antenna. See Fig. 2-1. As a general rule, longer is
better as far as the horizontal length is concerned, with the
lead-in kept as short as possible. The minimum length that
will pick up a signal other than from a local station is usually
some 40 to 50 feet, and a decided improvement results if it
can be extended to 100 feet or more.

A shows direction of maximum sensitivity

for incoming signals
^A
♦

lead-in

to receiver input

Figure 2-1: Diagram of an inverted L antenna and lead-in,

giving the direction of maximum sensitivity.

Ideally, the antenna should be 20 to 30 or more feet high, and

kept well away from buildings, trees, power lines, and any
other large improvements on the land. However, this is not
always practical except for the lucky few that live on a large

48
open tract of ground. Usually a compromise layout must be
made for a city lot, or where other buildings, hills, or trees
dominate the landscape, and this compromise extends to both
length and height. Fortunately, there are some alternatives
that help the situation. Sometimes a long, low antenna works
better than a short, high one. Changing the lead-in end or the
antenna direction can make a big difference. Of course, all
of these options depend upon antenna site limitations.

Another type of antenna with some of the characteristics of

the inverted L is the vertical antenna. Its big advantage is
that it can be erected where there is practically no room for
anything else. Its big disadvantage is that some kind of mast
is needed to hold it, unless there is an appropriate existing
structure available. Such an antenna must be at least 30 to 40
feet in vertical length to have characteristics comparable with
the shortest usable inverted L. A wire running down the side
of a building of this height and a few feet from the wall is
one possibility. Dropping the wire from a heavy horizontal
line between two buildings is another. A mast of adequate
height is always a solution, but it does require considerable
construction unless a kit is purchased. These are available
from suppliers that specialize in a full line of amateur radio
equipment, because they are used for transmitter antenna
masts.

The chief difference between a vertical and an inverted L

antenna is that the vertical antenna is omnidirectional. That
is, it is equally sensitive to signals from all points of the
compass. This may or may not be an advantage, depending
upon individual requirements. The lead-in can be very short
if the lower end of the vertical is not too high and if it can be
located close to the receiver. The unique vertical layout re­
quirements often make this possible.

49
The antennas described so far have the lead-in attached at
one end. The name "end-fed antenna" is sometimes applied
to this category, although the term is more appropriate for
transmitter antennas. End attachment is not a necessity,
however, because the lead-in can also be placed at the mid­
point. The simplest arrangement of this type is the T
antenna.

A T antenna, shown in Figure 2-2, is bi-directional, being

equally sensitive to signals from either direction along the
wire. It is a very advantageous type for problem installations
requiring receiver location below the antenna, rather than
near or beyond one end. This shortens the lead-in, which can
make a considerable difference in performance.

An L or vertical antenna cut to one-quarter wavelength (for

the frequency of reception) will be resonant. In fact all an­
tennas cut to one-quarter will be resonant. The advantage of
doing this, if space allows, is that the antenna can then be
tuned with the receiver for the reception of shorter wave­
lengths (higher frequencies) too. Hence, the ideal L or ver­
tical would be cut to one-quarter wavelength for the lowest
listening frequency desired. For example, by cutting the an­
tenna for the bottom of the AM band, it is possible to tune it
throughout the band. The T antenna, on the other hand,
would have to be doubled in length; each side must be one-
quarter wavelenth long. This is a serious disadvantage if the
T is to be used for the broadcast band - and you don't have
much space!

Another type of antenna that is very similar to the T in many

respects is the dipole. Instead of a single connection at the
mid-length, the two halves of the total length are separated

50
by an insulator, and the output of each half is fed separately
to the receiver. This type has been seldom used with crystal
sets, although it is highly favored for other applications.

h % %

receiver input

Figure 2-2: A typical T antenna.

All of the antennas previously discussed are broadcast band

tunable at the receiver, provided that their length is some­
thing like 40 to 50 feet or more. Antenna tuning is a highly
desirable feature that will be discussed later in detail. If you
are interested in a highly selective set that still retains rea­
sonable sensitivity, it is a definite advantage to have it.
Nevertheless, circumstances sometimes prohibit an outside
antenna or even a long single wire in the attic or down a
hallway. There is one other tunable type that can be used in
a room, and that is the loop antenna.

In situations where a straight wire antenna cannot be used,

then a nonresonant antenna can be considered as a last resort.
The advantage of antenna tuning is lost, but it does work in
most cases, sometimes all too well if there is a strong local
station. A broad band of interference with other stations re­
sults. A nonresonant antenna consists of one or more wires
strung around the walls of a room, the ceiling, or in an attic,

51
with bends and corners as necessary to increase the total
length of wire. Even electrical wiring can be used, but re­
member to install a series blocking capacitor and current
limiting resistor for safety.

All types of outside antennas share common requirements, in

terms of construction, in order to withstand continuous expo­
sure to the elements while functioning properly at the same
time. Practical problems of choosing the proper wire, use of
insulators, and installation technique are discussed in this
and a following section.

Antenna wire should be of large diameter, both for good

electrical conductivity and for strength. The best choice for
lengths up to a hundred feet or so is No. 7-22 insulated cop­
per. This consists of seven twisted strands of No. 22
American Wire gauge (AWG) copper wire with overall
rubberized or plastic insulation. The insulation serves as
protection from the outside atmosphere. Bare wire can be
used, but it has a lifetime that depends upon the environment.
In the dry southwest it may last for years in rural areas. Un­
der harsh marine air conditions its life may be measured in
months.

The stranded wire recommended here has good strength

characteristics to withstand wind and icing conditions. Solid
copper can be substituted, but large diameter wire must be
used to provide the same strength as the equivalent stranded
type. No. 14 AWG single conductor copper may be used for
the 100 foot category, and No. 12 AWG for 200 feet. The
breaking strengths are 213.5 and 337.0 lbs., respectively. A
copper coated steel wire, trade named Copperweld, can be
used for very long lengths of several hundred feet. Its effec­
tive conductivity is near that of solid copper while it has a

52
much greater tensile strength. No. 14 AWG Copperweld,
with 30% copper, has a breaking strength of 440 lbs., over
twice that of the equivalent 14 AWG solid copper.

Copperweld and similar copper clad products conduct high

frequency alternating current almost as well as solid copper.
This is due to a phenomenon called skin effect, where current
at radio frequencies is high at or near the surface of a con­
ductor and decreases toward the center. Skin effect permits
the substitution of copper clad steel for pure copper without
significant losses at radio frequencies. Copperweld is avail­
able in a range of sizes for both stranded and solid core wire.

Stranded wire is less susceptible to damage during shipping

and handling than solid core, and heavy insulation decreases
the likelihood of damage even further. All wire should be
inspected for defects before installation, and this is especially
important for solid core bare or enameled wire. That type is
particularly susceptible to kinks. A kink seriously cuts the
strength of wire, for both breaking strength and fatigue. Fa­
tigue strength is the resistance to repeated application and
removal of stresses caused here by wind or other sources of
intermittent loading. A kink can be removed in only one
way: by cutting it out and then splicing the two wire ends
back together. Solid copper wire that has less damage
caused by bends or loops can be salvaged by stretching it to
the yield point, which straightens it out. The same is true of
high strength wire, but that same property makes it more
difficult to stretch by simply pulling one end. Leverage of
some sort may be necessary.

Installing the wire requires that it be pulled taut enough to

avoid excessive sag at the midpoint, a condition typical of
any horizontal wire. Even a vertical antenna wire must be

53
sufficiently pre-stressed to prevent excessive movement in
the wind. The loading force that results can be considerable,
and it must be continually withstood during the lifetime of
the antenna. In addition, there are the temporary forces
caused by wind and ice. A very long antenna may require a
spring-loaded termination at one end to compensate for tem­
perature expansion and contraction of the wire. Posts,
frames, masts, and their attachments may be quite simple for
a short antenna, but can become elaborate for arrays and long
wire installations. These can be design projects in them­
selves and are not treated here.

Insulators are required at each end of an antenna, regardless

of what type of supports are used. These are available in
either glass or glazed porcelain and are impervious to mois­
ture. Types that absorb moisture should not be used. Ribbed
construction for some of the insulators increases the surface
path between the two ends, so that there is less leakage cur­
rent when dirt and moisture buildup occur. A simple rod
with a hole in each end works about as well for an antenna
wire insulator except under worst case conditions. Plastic
insulators are also available, but their use is somewhat lim­
ited by the relatively low tensile strength of plastics.

Ground

A good ground is essential for high sensitivity reception.

The best choice for a ground connection is a metal cold water
pipe, and if there is one inside the building and close to the
receiver, the need for a second feed-through to the outside is
eliminated. In such a case, an outside ground must still be
provided in a separate circuit with a lightning arrestor for
safety considerations. This will be discussed in the next sec­
tion. A typical ground system using one outside pipe is

54
shown in Fig. 2-3. The second feed-through is necessary
when no inside ground is available.

A hot water pipe makes a questionable ground. This is be­

cause the plumbing at the hot water tank may not provide an
all metal path back to the cold water inlet. A gas pipe should
not be used as a ground because of safety considerations,
even with presumably adequate protection from lightning or
electrical circuitry. New housing may come with some of
the piping in PVC, in which case there may be no inside
ground at all. A problem can also arise if the receiver is lo­
cated in a room well away from metal piping. In these
situations, some other grounding method must be used.

insulated lead-in fittings

antenna lead-in / use feed-throueh porcelain or
/ equivalent for bare wire leads
ground lead-in
A ant
lightning arrestor
? receiver
(must be located gnd
outside of building)
14 AWQ or heavieL
copper ground wire
pipe clamp

water pipe ->

Figure 2-3: A ground system using an outside pipe and

lightning arrestor.

A simple outside ground that works in moist soils is a one

half inch or larger diameter metal pipe driven some three feet

55
or more into the ground. Filling the pipe with water may
help where the soil tends to dry out. Reception may improve
if several pipes are driven into the ground at intervals of a
few feet, or in a cluster, and are connected together. Another
possibility is a metal plate buried in the ground. In dry cli­
mates, it may be necessary to supply several square feet of
plate, at a depth at least equal to the principal plate dimen­
sion.

Lightning Protection

A lightning strike, as we all know, can injure or kill as well

as do property damage. Protection of the antenna itself from
lightning is impossible, but a lightning arrestor is a useful
device for preventing the high current surge produced by a
lightning strike from entering a building. The system is
shown in Fig. 2-4. The lightning arrestor in the figure nor­
mally acts like an open switch, but when lightning strikes the
antenna, arcing occurs, and it becomes a short circuit across
the lead-ins to the building. The current surge from the an­
tenna is diverted to the ground connection outside the
building. This is the reason an outside ground is necessary
even when the receiver is grounded separately inside.

The operation of the lightning arrestor depends upon two

heavy metal electrodes spaced a few thousandths of an inch
apart. The resulting air gap is the "open switch" referred to
above. RF signal voltages and normal static are far less than
the breakdown voltage across the air gap. The situation
changes drastically when the antenna is struck by lightning.
The voltage across the gap abruptly rises to a high value,
reaching the breakdown voltage. The air in the air gap re­
gion ionizes, and a very low resistance path is established
between the electrodes. This resistance is much less than the

56
circuit resistance through the building and back, so the high
current surge produced by the lightning is shunted directly to
the external ground.

Suitable lightning arrestors are available at many hardware

stores and electrical supply houses. It is also possible to
construct one, using a spark plug re-set for about a one mil­
limeter air gap. The narrow air gap should be protected by a
plastic cup or some similar cover, fastened to the end of the
plug. A pipe clamp can be used to connect to the opposite
end. This is an arrestor in its simplest form, and a more ele­
gant version can be constructed if desired.

insulator antenna insulator

anchor plate I
lead-in
stand-off
insulator

feed-through
insulators
m md antenna connection
to lightning arrestor

lead to ground lightning arrestor

from receiver grounded at pipe

pipe for receiver ground

Figure 2-4: General arrangement for an inverted L antenna,

with lead-in, ground for receiver, and lightning protection for
building interior.

The preceding sections cover the complete set of components

needed for an antenna-ground system. This brings the signal
to the tuned circuits of the receiver, where the wanted signal
is separated from unwanted signals and most of the noise.

57
Signal Selection

The previous sections have covered the antenna, ground and

lightning protection aspects of crystal set receivers. The task
of separating the signal of just one station from many and
delivering the audio signal to the headphones are discussed
below.

The voltage output from the antenna consists of the sum of

the RF signals and noise that may be present. It is desirable
to select a particular signal as effectively as possible and then
apply that one RF voltage to the crystal detector. There are
three ways in which to do this:

1) Pass through the wanted signal in terms of an

increased voltage level.

2) Reject the unwanted signals in terms of decreased

voltage levels.

3) Use a combination of both methods.

The first alternative is the method usually used. This is ac­

complished with an LC circuit—a combination of a coil and
a capacitor—which is tuned to the frequency of the desired
signal. The signal voltage at the resonant frequency is
greatly increased, so that the wanted signal is delivered to the
detector at a much higher level than noise and signals at
other frequencies. However, noise at the resonant frequency
cannot be separated from the signal, and it also reaches the
detector.

The rejection method is used in a few special cases and also

employs a tuned LC circuit. For this situation, the resonant

58
frequency of one signal is rejected while signals and noise at
other frequencies are passed. This type of circuit is com­
monly called a wave trap. It is used with at least one other
tuned circuit to combine both methods of signal selection,
producing one signal at the output.

More elaborate LC band pass filters and band rejection filters

are seldom used in crystal sets because there is too much
attenuation of the desired signal. This more than offsets their
superior filter characteristics. Their application is in systems
where amplification can easily compensate for signal loss.
Resistance capacitance (RC) filters are never used at RF fre­
quencies because of excessive signal losses. The practical
methods of signal selection are discussed in detail below, as
applied to each section of the crystal set, beginning at the
antenna.

Antenna Tuning

Modem receiver design does not apply to the crystal set, as

far as antenna tuning is concerned. Ganged tuning with one
control was considered an essential feature for the home re­
ceivers built since the 1920's (superheterodynes). Such
tuning was designed to optimize operation of the mixer and
intermediate frequency circuits. Getting the last bit of radio
energy out of the antenna was not the goal. To optimize the
crystal set, however, the antenna and LC circuits must be
tuned separately. There is an impressive improvement in
crystal set performance when the antenna is tuned.

The most commonly used tuned antenna is the inverted L,

discussed earlier. Both inverted L and vertical antennas are
self-resonant at a wavelength which is roughly four times the
length of the antenna. The relation is not exact because,

59
among other things, the capacity to ground and the length of
the lead-in also affect the resonance wavelength, as was
shown earlier. The shortest antenna which is self-resonant in
the broadcast band is about 150 feet in length, corresponding
to 1650 Khz, and it is nonresonant at lower frequencies.

The antenna system resonant frequency decreases when the

antenna coil of the receiver is connected between the antenna
and ground. A variable inductance can be used to tune the
antenna to lower frequencies, but it is usually more conven­
ient to add a variable capacitor in series with the antenna coil
and use it for tuning. A conventional 365 pF variable capaci­
tor with a 250 pH coil can be used with a long antenna to
tune through the entire broadcast band. A practical circuit
for series-tuned antennas is shown in Fig. 2-5. Shorter anten­
nas require additional inductance, capacitance, or both, to
tune below about 800 to 1000 kHz, and an antenna of less
than 40 feet is usually unsatisfactory for antenna series tun-
mg. These very short antennas require that a tuning
capacitor in the circuit be connected in parallel with the
antenna coil instead of in series, so that the antenna coil itself
is tuned. This is shown in Fig. 2-6. A practical arrangement
for selecting either arrangement is given in Fig. 2-7.

In principle, the variable capacitors in Fig. 2-5 can be located

at either the antenna or ground side of the coil, and this is
true in practice for crystal sets which are not selective. How­
ever, a set that is very selective because of high circuit
efficiency (high Q) requires that either the variable capacitor
be on the ground side or that it be thoroughly isolated from
the front panel controls. This is because stray capacitance to
ground changes when a hand is brought near the panel,
which shifts the antenna circuit resonant frequency. This
hand capacity effect becomes very noticeable and annoying

60
with sharply tuned circuits that are above ground potential.
The easiest way to avoid the problem is to ground the capaci­
tor rotor, as is done in Fig. 2-5.

v M=mutual inductance
coupling to detector circuit
->

antenna coil to detector circuit

<y*°o

365pf variable capacitor

additional for use with short
and medium-length antennas
(365pf or more]

Figure: 2-5: Practical circuit for antenna tuning.

61
 V
M=mutual inductance coupling
to detector circuit

365pf variable
capacitor

Figure 2-6: Example of antenna coil tuning with a short

antenna.

V A good antenna tunes very sharply

when coupling to the rest of the set is
not too tight, and a narrow band, high
amplitude signal is produced at reso­
■o i
nance. This is why antenna tuning is
so important. This is also why a
straight wire antenna with a short lead-
in is so desirable. Tuning also works
well with more than one wire con­
nected in parallel, provided that the
o---- wires are strung parallel and are of
equal length, such as in an inverted L
array. One or more wires strung ir­
Figure 2-7: Switching regularly around several comers may
circuit for selecting
series or parallel result in a loud signal, but they are
tuning of antenna. untunable.

62
Detector Circuit Coupling and Tuning

Some of the simplest crystal sets place the detector directly

across the antenna coil, as in Fig. 2-8. Either Ca or is
used for tuning. When both are present they interact, but
there is sometimes an advantage for such a system because
of its flexibility. In any case, the disadvantage of the Fig. 2-
8 circuit stems from the fact that the detector and headphones
series circuit across the coil has a relatively low impedance
and loads the coil, decreasing the Q. Lowered Q means

v i
#cd

Ca

Figure 2-8: Simple form of crystal set with detector directly

coupled to antenna coil.

lower circuit efficiency and broader tuning for both the an­
tenna and detector. Overall selectivity is often poor, even
when both capacitors are used and set for best operation.

63
Hand capacity with Ca on the antenna side, rather than as
shown, is not a problem because of this lack of selectivity.
One method of decreasing loading to improve selectivity is
to tap the detector coil and load only a portion of it. The use
of a tapped detector coil is discussed later.

A more complicated but more efficient system appears in

Fig. 2-9. Here the detector is coupled by variable mutual in­
ductance to the antenna coil, and this permits some choice
between close coupling for improved signal output (high
sensitivity) versus loose coupling for improved selectivity at

v M

4
cd
Ld

7FC*

Figure: 2-9: Mutual inductance coupling between antenna

coil and detector coil.

the expense of low sensitivity. As before, a tapped coil can

be used to decrease detector loading. Standard values of in­
ductance and capacitance can be used for Ld and Cd, and

64
changing the tap location does not greatly affect the setting
of Cd to maintain a given resonant frequency. Fig. 2-9 can
be used as the basis for some excellent crystal set designs.

Special coils are a desirable alternative to commercially

available products. In general, and Cd may be chosen in
terms of the resonant frequency formula,

1
/,=
2n^LdC(l

where fr is the minimum resonant frequency of the tuned

band in Hertz, is the inductance in Henrys, and Cd is the
maximum value of the variable capacitor in Farads. In more
practical form the basic relation is

159,155
fr =

where the units of measurement are

[Cd ] = picofarad
[fr] = kilohertz
[Ld ] = microhenry

Another method of coupling the antenna coil to the detector

is by means of small capacitors, as shown in Fig. 2-10, and is
called capacitive coupling. This method was used to some
extent in the early days of radio, but later lost popularity in
favor of the inductive coupling methods described in preced­
ing paragraphs. There seems to be no particular advantage in
this method to offset the requirement for additional parts.
Indeed, at least one early day textbook stated that capacitive

65
coupling was less selective than the inductive method. One
or both coupling capacitors can be variable, and the return
side capacitor has been omitted in some versions of the cir­
cuit without notable differences in performance. In that case,
the return circuit depends upon stray capacitance. It is nec­
essary to reduce mutual inductance between the two coils as
much as possible here, or operating peculiarities may result.

Cc
V J-
4
La
cd
Cc

Ca

Figure 2-10: Capacitance coupling between antenna coil and

detector coil.

Tuned Transfer Circuits

The possibility exists of improving the performance of such

circuits as the one in Fig. 2-9. The approach is to change the
way energy is transferred from the antenna to the detector to
an arrangement where a transferring circuit is tuned. In the­
ory, only the desired signal energy is carried over from the

66
antenna coil to the detector. At least two types of circuits
have been employed, both dating back many years.

The most straightforward and easy to understand method is

tuned link coupling, one version of which is shown in Fig. 2-
11. Here La and are isolated from each other as much as
possible. Energy is transferred from one to the other by
means of Ln, Ln, and C(. In another form of this circuit,
L(l, L(2, and C( are all connected in parallel. Variations are
possible, such as grounding one or both floating circuits.
Detailed performance information on link coupling for crys­
tal sets is lacking, but it has been much used in the past. One
of the most highly regarded receivers of the World War I
period was the American Marconi Type 107A, and it used
tuned link coupling with a crystal detector in one of the sys­
tem arrangements.

V Q

4
<=D Lp L&< Ld ^-Cd
La

Ca

Figure 2-11: Link coupling between antenna and detector

circuits.

The second type of circuit appears in Fig. 2-12. Good results

have been reported for some variations on this basic ar-

67
rangement, where energy from the antenna coil is transferred
by a single tuned coil located between the antenna and detec­
tor coils (all in-line). The Q of the system as a whole is very
high when good coils are used, and selectivity can be surpris­
ingly high without undue loss of sensitivity. One of the
drawbacks of this approach is related to that high Q circuitry.
Such systems are loosely coupled by necessity, and so be­
come extraordinarily sensitive to mutual inductive and stray
capacitive effects between the components and to the sur­
roundings. The operating peculiarities which sometimes
result can only be described as bizarre.

La

4
<=>Ld cd

ca

Figure 2-12: Transfer coupling between antenna and detector

circuits using a third tuned circuit.

68
i
Wave Traps

Wave trap circuits were quite popular at one time, when it

was difficult to design sharply tuned receivers. They still
have their applications with crystal sets, where selectivity is
often a problem. A wave trap is a simple form of filter and
can be used in the same way as band pass or band rejection
filters are used. An acceptor wave trap passes its resonant
frequency while a rejector wave trap rejects that frequency.

A form of acceptor wave trap which improves the perform­

ance of a simple crystal set, such as the one in Fig. 2-8, is
shown in Fig. 2-13. Here the trap LtCt is coupled by mutual
inductance to the antenna coil to boost the output at the reso­
nant frequency. This has sometimes been called absorption
tuning. Depending upon the amount of coupling M, there
may be noticeable interaction between C* and the other ca­
pacitors when tuning the set. In principle, this approach can
be used with any tunable circuit.

wave trap
v
i
l-d — ^

Figure 2-13: Acceptor type wave trap for increasing signal

output to detector.

69
The rejector wave trap is most commonly used in the an­
tenna-ground circuit to tune out a strong unwanted signal
which is interfering with reception. This arrangement is
shown in Fig. 2-14A. The version in 2-14B is often prefer­
able because the amount of coupling can be adjusted for a
particular set of operating conditions. Some other wave trap
locations are shown in Fig. 2-15.

v V

U§ Cr l,| k
4 4
ILA<=>
° LA,
<=>

7-C* CA

(A) (B)

Figure 2-14: Rejector wave trap designs for antenna circuit

applications.

Wave traps all have a common disadvantage; they interact

with the associated tuned circuits. This is not a major prob­
lem with acceptor type circuits such as Fig. 2-13, but rejector
circuits seriously affect the tuned circuit characteristics. The
concept of tuning to a weak distant station and then tuning a
rejector wave trap to eliminate a second interfering station is
an impossible dream. The resulting change in the system
resonant frequency is so great that the desired station is no

70
longer in tune, and if it is too close to the wave trap resonant
frequency it may be lost altogether.

The only practical usage is to preset the trap at the frequency

of an interfering station and leave it there while operating in
that general portion of the band. Also, a rejector trap is use­
ful only for strong local signals that have a broad bandwidth,
because the trap itself is not selective enough to discriminate
between two signals of comparable amplitude and close to­
gether in frequency. In many situations it is better to
optimize set selectivity than to depend upon wave traps, but
special cases do arise where a rejector trap may be helpful.

V wave trap
V L,
GW0
4 !SLd cd

wave trap
^ L,
V nsrooot
*

4 Ld cd
t>
Figure 2-15: Rejector wave traps in detector circuits.

71
Signal Detection

The signal that reaches the detector is still a radio frequency

voltage, regardless of the amount of signal processing be­
tween the antenna and the detector input. It consists of an
RF carrier modulated by an audio waveform. Fig. 2-16
shows a typical signal at the detector input. Here, the enve­
lope is the modulating audio signal, and the positive and
negative components of the envelope are 180° out of phase.
Any device sensitive only to audio frequencies would display
zero output if connected at this point, since the two envelope
components cancel at any given time. A peak-to-peak RF
voltmeter would show either a constant or fluctuating volt­
age, depending on the signal and meter inertia, unless the
carrier were unmodulated. In that case, a constant RF volt­
age output would result.

Radio Frequency Carrier Positive Envelope

0 TT
M
1/ v
-V

Figure 2-16: Radio frequency signal at input of diode

detector.

72
The Detection Process

Crystal detectors fall into the diode detector category, and a

characteristic common to all diodes is the property of rectifi­
cation. By definition, a rectifier is a device which passes
electrical current more easily in one direction than the other.
An ideal rectifier has zero resistance to current flow in its
forward direction, and infinite resistance in the opposite
(reverse) direction. Therefore, with an ideal rectifier, the
input voltage shown in Fig. 2-16 produces zero current
through an output-side load whenever the input voltage is
negative. The output voltage across the load must appear as
in Fig. 2-17 since there is current flow whenever the input
voltage is positive, with a corresponding voltage drop across
the load.

Positive Envelope

XX

0 I
\ Positive Half of
RF Carrier

-V

Figure 2-17: Idealized diode detector output after half-wave

rectification.

A test circuit for measuring diode characteristics is given in

Fig. 2-18. It is convenient to use a commercial model signal
generator with a RF output which can be either modulated by
an audio sine wave or unmodulated. Meter Vjn may be

73
available with the generator to display the RF output voltage.
The characteristics of the crystal are determined when the
switch is closed to shunt out the load impedance. In that
case, the output current produced by unmodulated input volt­
age Vjn is measured by AC millimeter i. Behavior under
various output loads is determined by opening the switch and
reading the output voltage. Input voltage is kept in the order
of one volt or less for the type of measurements which are
applicable here.

Forward voltage connection to crystal

RF Signal Qenerator depends on material

o 4
o
V,n

0 \ Load
<r
Figure 2-18: Test circuit for measurement of diode
characteristics.

The output current waveforms for ideal and real solid state
diodes are shown in Fig. 2-19, with zero impedance load.
The ideal detector permits no negative current flow. How­
ever, even the best real diodes allow some flow in the reverse
direction, because they do not fully rectify the current. That
is why there is a small negative current flow in Fig. 2-19B.
This current subtracts from the forward current, on an aver­
aged basis, so that the net audio output current i is lower than
for an ideal diode. Nevertheless, a good diode will show a
substantial net forward current output.

Direct current measurements also provide a great deal of in­

formation about diode detectors, In fact, early day
investigations of crystal detector properties depended upon
this type of measurement. The circuit, shown in Fig. 2-20, is

74
quite similar to Fig. 2-18, but a variable DC power supply is
used, with some convenient provision for reversing the out­
put polarity. Measurements are made with zero impedance
loading. Commercial diodes are evaluated over a wide range
of positive (forward) and negative (reverse) voltages to fully
determine their characteristics, but here the necessary DC
input voltage ranges over only a volt or so in magnitude,
with corresponding current output in the order of milliam-
peres or less. With this arrangement, the voltage-current
curves for ideal and real diodes are given in Fig. 2-21.

+i

Zero Current -►

-i
Time Time
A) Ideal Detector B) Real Detector

Figure 2-19: Output current characteristics of ideal and real

diode detectors with zero impedance loads.

The ideal detector, in Fig. 2-21 A, conducts perfectly when

the input voltage is positive, because it has zero internal re­
sistance in the forward direction. No current is passed when
the input voltage is in the reverse direction because the inter­
nal resistance becomes infinite. Should an ideal diode be
measured with a load resistance at its output, the output cur­
rent becomes iout = Vjn /(load resistance) in the forward
direction and zero in the reverse direction.

75
variable voltage
DC power supply
+ I

Figure 2-20: Test circuit for measurement of diode DC

characteristics.

A real diode has a finite internal resistance, which prevents it

from conducting perfectly in the forward direction. Conse­
quently, i0ut rises as shown when Vin is increased in the
forward direction, and is given by iout = Vjn/(intemal resis­
tance + load resistance). The internal resistance does not
remain constant as the voltage increases, so the forward cur­
rent versus voltage curve is not a straight line, although it
usually becomes approximately straight at angle 0 in the

T~_ forward current

forward cu rent
slope of forward
g g -s current curve
|o |o 0;
£■ I & /
8 reverse cun *nt 8 reverse currcr t

0 0
input voltage input voltage
A) ideal diode B) real diode

Figure 2-21: Current-voltage curves for ideal and real solid

state diodes.

76
figure as the voltage is increased. In the reverse direction,
the diode resistance is much higher but still finite, so a low,
gradually increasing reverse current results as the negative
voltage is increased in magnitude.

The curves in Fig. 2-21 are applicable at RF frequencies as

well as for DC. The net current which determines the
modulating signal output is the difference between the for­
ward and reverse currents for a given peak to peak RF volt­
age. The forward current results from the positive peak
voltage. The reverse current results from the negative peak
voltage. This was shown for RF voltages in Fig. 2-19B,
where the presence of the small negative RF voltage actually
decreases the modulating signal output. The difference be­
tween the two peak current magnitudes in 2-2IB, not their
sum, is the measure of how large the audio output may be.

The difference between the voltage envelopes in Fig. 2-19B

is the maximum available audio output from the detector.
This net voltage is readily convertible to audible sound by
means of earphones, without significant additional signal
processing, as is discussed in the next section. First, how­
ever, it is necessary to consider the current-voltage curve in
further detail, and to describe a property that can be used in
some cases to increase the available signal output.

The internal resistance of a crystal decreases as the forward

voltage across it increases. This was mentioned above, and it
is another demonstration that active substances do not follow
Ohm's Law. As a result, more current is produced when the
resistance is lower; that is, sensitivity is higher when the
slope 0 is greater. On the other hand, if a point near zero
voltage were chosen, where the slope is less, sensitivity is
greatly reduced. By imposing a DC voltage on the crystal,

77
even though the RF signal may be small, we can operate the
diode where 9 is larger (current is higher). Hence, sensitiv­
ity to a weak RF signal is increased. One way to do this is
shown in Fig. 2-22 and is called biasing the crystal.

Current-voltage curves differ depending upon the material of

the diode. The best operating point 0 in many cases is so
close to zero that no bias is necessary. In other instances,
over a volt of bias is needed for best sensitivity. Carborun­
dum is a classic example where several volts of bias may be
necessary.

adjustable low
voltage source
(+) (-)
bias voltage is q Q
introduced across H-
to antenna dropping resistor:
and ground
4
RF bypass capacitor
shunts resistor

Figure 2-22: Method of biasing a crystal detector by

introducing a constant voltage into the detector circuit
voltage loop.

The Audio Output

The electrical and mechanical characteristics of high imped­

ance magnetic diaphragm earphones are usually sufficient to
provide a satisfactory output load for a crystal detector. A
typical output schematic and its equivalent circuit are shown

78
in Fig. 2-23. The inductance of the coils, wound on mag­
netic cores, act as chokes to stop flow of RF current. The
self capacitance of the coils is usually sufficient to store RF
energy during each half cycle of the carrier voltage. Some­
times a capacitor C may be added as shown in the figure to
improve upon this situation. In any event, the net result is a
filtering action which permits passage of an average signal
level that closely approximates the RF envelope. See Fig. 2-
24. The diaphragms of the earphones respond readily to this
signal to produce the required sound. At the same time, they
are mechanically insensitive to any RF ripple still present at
the output, because they cannot vibrate at such high ffe-
quencies. As a result of both electrical filtering and
mechanical inertia, the output of the detector is converted to
sound.

4--- r—O-
1
RF
resistance
c# self in­
Qen
ductance
O
self capacitance
j equivalent circuit
i or earphones

AJ Crystal Detector and Output Circuit B) Equivalent circuit of Crystal Detector

and ouput circuit

Figure 2-23: Crystal set output circuit.

The high impedance magnetic diaphragm earphones de­

scribed earlier are the most commonly used type for crystal
sets. Other types are available and some have been used.
These may or may not perform well and in some cases re­
quire additional circuitry to provide a good impedance match
to the output of the receiver. The most often used of these
are piezoelectric earphones, often called crystal earphones.
These depend upon the expansion and contraction of a mate-

79
rial which changes its dimensions when an electric field is
applied across it. This change in dimensions drives a dia­
phragm to produce sound.

filtered voltage
forward voltage
+V envelope

RF voltage
0
reverse voltage
envelope

Figure 2-24: Output voltage at earphones.

The impedance across the input terminals is significantly

higher than even the best quality magnetic diaphragm units,
so there is less loading of the detector circuit. However, re­
sults are satisfactory only with better quality units such as the
Clevite "Brush" type with impedances as high as 200,000
Ohms at 1,000 Hz. In any case, a shunting resistor is neces­
sary across the earphones to prevent DC voltage build-up,
because the impedance of the earphones is almost entirely
capacitive. Cheaper crystal phones require a lower shunting
resistance to decrease distortion and improve bass response.
This increases detector circuit loading and tends to eliminate
the advantage of using them.

Loudspeaker operation of a crystal set is not usually practi­

cal, since the available power is so small. There are a few
exceptional cases, often deriving from the fact that there is a

80
 high power transmitter nearby, and that the receiver has been
designed for high sensitivity. Results have been reported
when a small permanent magnet speaker is coupled by an
impedance matching transformer to the detector output.
Better performance is likely by converting an earphone to a
speaker by adding a conical hom at the ear piece; one of the
early day hom speakers using that same approach may be
employed. Usually, such a system is not very practical and
is treated as an oddity or stunt.

BIBLIOGRAPHY

Allied Radio Corporation Catalog 140, 1955.

American Radio Relay League, The ARRL Handbook for

Radio Amateurs. 70th Edition, 1993.

Anderson, Philip N., The Crystal Set Handbook. The Xtal

Set Society, Inc., Lawrence, Kansas, 1994.

Bucher, Elmer E., Practical Wireless Telegraphy. Wireless

Press, Inc., New York, 1917.

Chemical Rubber Publishing Co., Handbook of Chemistry

and Physics 35th Edition. (Also both earlier and later
editions).

Edwards, K. E., Radios that Work for Free. Hope and Allen
Publishing Co., Grants Pass, Oregon, 1977.
!

Ghirardi, Alfred A., Radio Physics Course. 2nd Edition,

Murray Hill Books, Inc., New York, 1932.

81
Green, Charles, All About Crystal Sets. Allabout Books,
Fremont, CA, 1984.

Grover, Frederick W., Inductance Calculations. D. Van

Nostrand, 1946. Reprinted by Instrument Society of
America, Research Triangle Park, North Carolina, 1973.

Pierce, George W., Principles of Wireless Telegraphy.

McGraw Hill Book Co., Inc., New York, 1910.

Robertson - Cataract Electric Co., Radio Equipment and

Supplies. Catalogue No. 22, 1922-1923. Reprinted by The
Vestal Press, Ltd., Vestal, New York.

Rosen, Marvin and Phil Anderson, communication in The

Xtal Set Society Newsletter. Vol. 4, No. 1, Jan 1, 1994, p.5.
(concerning suggested use of a transistor without external
power).

Terman, Frederick Emmons, Radio Engineering. 2nd

Edition, McGraw Hill Book Co., Inc., New York, 1937.

Thackeray, Desmand, The Carborundum Story. The Old

Timer's Bulletin. Vol. 27, No. 3, Nov. 1BL6, p.10.

82
 V CHAPTER 3
CRYSTAL SET DESIGN
L1
+
L2 Introduction
C O
C5j >C=D
There was a time during the early

I days of radio when crystal set

design was part of the main­
stream technology that we now call electronic engineering.
The use of crystal detectors and several other detection
methods evolved at about the same time, using similar cir­
cuits and components. In fact, some of the receivers of that
period had provisions for two or more detector types on an
interchangeable basis. Significant improvement of the first
crude vacuum tubes changed all of this, making the crystal
set obsolete, first as a primary method of signal reception,
and, after many years, even for standby or emergency use.
After the early 1920's, design methods diverged from crystal
set era technology as electronics developed along other
paths.

Present day electronic engineering texts cover crystal set de­

sign only where by coincidence there is still some common
ground. In fact, there are now problems in crystal set con­
struction which did not exist during the early decades of this
century. This is because some modem components do not
I work well in crystal receivers, and early day versions are no
longer produced. Situations arise where it is necessary to
either modify a modem part, or build one on a do-it-yourself
basis, all to provide something that could have been bought
off the shelf before 1930 or 1940.

This chapter gives the information needed to design radios

using crystal detectors, and there are several approaches to

83
this process. Among other possibilities, the designer may
want to surpass present day performance with something
new, reconstruct a circuit given only partial information, or
analyze an early day receiver to construct a modem replica.
These and other alternatives are covered in the section on
design criteria. It will be shown that some properties and
functions are inevitably emphasized at the expense of others,
and that there are decisions to be made in that respect before s
proceeding further.

A following section describes circuit options that are avail­

able, after the design goals have been established. There is
no one "best choice" or universal solution. Once a basic type
of receiver is chosen, there are numerous possibilities for
changing or adding features. These are covered in the final
section on circuits and systems with a proven track record for
good performance. Also worth consideration, and included,
are some complete receiver systems which are so unique in
one or more respects that they stand on their own. These are
presented for either direct application, or for comparison
with other options.

Design Criteria

Design depends upon application. This is just as true for

crystal sets as it is for anything else built by human hands.
Therefore, it is important to choose the desired application
requirements before proceeding with a particular design.
These usually include some of the following design parame­
ters:

♦Sensitivity — a "must" for long distance reception

♦Selectivity — for separation of interfering signals
♦Size — is compactness important?

84
 ♦Simplicity — straightforward construction,
convenient operation
♦Appearance — "traditional" or modem
♦Antenna — compatibility with what is available

The choice of criteria is not as easy as it first appears. A

desire to meet two or more requirements on the above list is
likely to produce conflicting conditions which can be met
only by compromise. The classic example of this problem is
the trade-off choice between sensitivity and selectivity. The
two properties are so completely interdependent that they
will be discussed together. Other criteria will be covered
separately, with some comment on interdependence where
necessary.

Sensitivity and Selectivity

Is your primary goal distant reception or local coverage

without interference? Superficially, this is a clear-cut choice
between a receiver optimized for best possible sensitivity or
for maximum selectivity. However, as a typical example,
many people live within twenty or thirty miles of several
broadcast stations, and just one of these can produce a wide
interference band on a low selectivity set. Any advantage of
high sensitivity is lost within this band. In fact, a high sensi­
tivity, low selectivity set is a useless artifact in metropolitan
areas with many transmitters, as far as distant reception is
concerned. On the other hand, emphasizing selectivity de­
creases sensitivity, making long distance reception less
likely, so it is necessary to consider the interdependence of
these properties.

There are mathematical formulas for selectivity and for

sensitivity, expressed as a ratio of output to input signals.

85
These depend upon each type of circuit. The formulas are
elaborate except for very simple circuits, with significant ap­
proximations required in some cases. For practical purposes
it is sufficient to state that a relationship between selectivity
and sensitivity almost always exists where increasing selec­
tivity results in decreased sensitivity and vice versa. The
simplest method of investigating that relationship is by direct
experiment.

Adjustment of component values and/or coupling for a given

receiver establishes a certain amount of trade-off between
selectivity and sensitivity. However, the relationship
changes when the circuit itself is changed, leading to a new
interdependence and a new choice of trade-offs. Adjusting
parts values or circuit coupling for a given receiver may not
be enough to obtain good results. It may be necessary to
change the entire receiver system. Some specific examples
are given below.

It is possible to design a selective tuner which will separate

local stations quite effectively, but the resulting signal levels
are low. This type of system has been used with audio am­
plification to improve the output level, but the result is not a
crystal set. In certain circumstances, a selective tuner driving
earphones without amplification has been advocated as an
excellent high fidelity AM receiver. Only local reception is
practical at best.

A multicomponent narrow band pass filter and amplification

may be used for these high fidelity applications, rather than
the tuning arrangements described in the previous chapter.
Although antenna tuning would improve reception, it is un­
necessary and omitted when amplification is used. Details
concerning band pass filter design are given in most electron-

86
ics textbooks and are not covered here. Signal attenuation of
the pass band signal through the filter is too great for crystal
set application.

The preferred approach for good crystal set selectivity is to

use a combination of antenna tuning with one or more of the
tuned circuits discussed in Chapter 2. The number of possi­
ble combinations of these is very large, especially when
including the variations of each circuit which can be consid­
ered.

As a specific design example, consider what must follow af­

ter selecting a system with moderate selectivity and
sensitivity for use outside of station-congested areas. A ba­
sic arrangement is shown in Fig. 3-1. This is a good system
from which to start, but by no means the only one. Some
comments can be made about this and similar circuits, based
upon accumulated empirical results. First, at least two tuned
circuits are necessary to reach a reasonable degree of selec­
tivity, unless a very insensitive receiver is acceptable.
Second, for application here, the "bigger is better" approach
to improved coil performance reaches the point of diminish­
ing returns for single layer air core solenoid diameters
greater than about 3 inches, and for wire sizes lower than 20
or 22 AWG, i.e., large diameter wires. Also, variocoupling
should be used as shown in the figure, at least on a prelimi­
nary basis, to determine the amount of antenna detector
coupling for the specific application. It may or may not be
desirable to provide for variable coupling as one of the con­
trols on the finished receiver. All of these limitations should
be considered as guidelines only, which may be exceeded
under some circumstances.

87
The component values shown in Fig. 3-1 for the antenna coil
and antenna tuning capacitor are representative for a medium
to long length inverted L antenna and broadcast band recep­
tion. One or both values may be changed for short or for
very long antennas. ;

v variable mutual
inductance coupling

o6 365 pf

4
250 ph
250 jib

365 pf

’:=r

Figure 3-1: One basic type of selective circuit.

Tapping the detector coil is important. It takes advantage of

a very unusual situation: both sensitivity and selectivity im­
prove over a limited range when the taps are used. First, the
signal output rises as the fraction of the detector coil length
that is loaded by the detector decreases, due to better imped­
ance matching between coil and detector circuit. At the same
time, selectivity improves as detector loading of the tuned
LC circuit decreases. At some point, sensitivity reaches a
maximum and then decreases as the switch is advanced fur­
ther, while selectivity continues to improve. Just where the
maximum occurs depends upon the rest of the circuitry, in
terms of component geometry and parts layout, as well as the

88
exact circuit design, because the system is influenced by
stray capacitances. It is not a very critical adjustment, but it
would be well to use at least five or six taps, if a continuous
sliding contact is not chosen as an alternative.

Tapped windings or sliding contacts were a normal feature of

early day receivers, but they fell out of favor in more modem
designs. Many crystal set circuits simply connect the detec­
tor to the upper end of the coil and good performance results,
but the improvement using a tap or slider at the detector coil
should more than compensate for the added system com­
plexity. It is used here to bring the selectivity up to a
moderate level without sacrifice of sensitivity.

Tapped windings or sliding contacts were also used in other

tuned circuits in early day receivers, and there was often
some benefit gained by varying the ratio of inductance to ca­
pacitance at a given frequency in a tuned circuit. This made
it possible to adjust for the most satisfactory operation. The
tuning of such a receiver demands real skill for effective use,
especially if there are two or three such circuits, because the
same signal can be tuned-in over a wide range of dial set­
tings. These options have not been used here because the
added benefit would be unlikely to make the inconvenience
worthwhile. Such an addition would be worth considering
for a more extreme design, to trade convenience in favor of a
better combination of selectivity-sensitivity.

The type of coil which is used is not too critical for the pre­
sent example, provided that it is adaptable for tapping or for
a sliding contact, and that it is a high Q design. Air core
single layer solenoids are the usual choice. Basket weave
coils are another good example, and a short single layer sole­
noid wound on a ferrite core works well.

89
Some care should be used in selecting other components
when considering selectivity and sensitivity. Most 350 or
365 picofarad air dielectric variable capacitors, modem or
early day, are satisfactory for the broadcast band. The 140
pF size is standard for short waves. However, modem sub­
miniature units that depend in part on a solid dielectric may
not qualify. The Q ratings of some of these are less than de­
sirable for an optimum choice.

Size

This is a parameter which may be all-important to some and

of no consequence to others. As an example, during the
early to mid 1920's, radio oriented publications gave consid­
erable attention to the "smallest" crystal set, and to tiny re­
ceivers housed in unusual containers. A typical case was a
crystal set built into an earphone shell, or built upon the
outside of a working earphone, which was used at its output.
Performance must have been poor, and the fad is hard to un­
derstand in view of the necessity for external antenna and
ground, which eliminated any possibility of being portable
while operating. It is evident that size is the dominant design
choice for replicating a set of this type.

A more contemporary application is the design of improved

miniature receivers as replicas of the little sets advertised for
many years as toys or novelties. Miniature versions of coils
and variable capacitors are available now that are adaptable
for such an application.

Size is also linked to some degree with sensitivity and selec­

tivity, because size affects coil Q, and to a lesser extent the Q
of variable capacitors. Small coils are inefficient. Further, if

90
a hand capacity problem arises, isolation of the offending
components may require a larger layout. Thus, if both size
and sensitivity-selectivity are important, a design trade-off is
necessary.

Simplicity

Straightforward circuitry, layout, and packaging become im­

portant when designing for production. This is undoubtedly
a factor in the case of the small businesses that advertise
crystal set kits and fully assembled receivers for sale. This
type of enterprise cannot afford to produce a receiver requir­
ing excessive parts and extra labor for their assembly.

Simplicity is also a factor that the beginner should consider.

It is wise to start with a simple project at the beginning of the
learning process, in terms of knowledge of electronics, con­
struction technique, and crystal set limitations.

If you wish to build a tried and true receiver design, rather

than experiment, then avoid elaborate high Q systems. High
Q, as applied to over all system efficiency, is quite possible
with some of the sets using several tuned circuits. This is in
spite of the fact that power is consumed in circuit losses and
at the earphones. However, high Q usually means increased
circuit complexity. Also, stray capacitances and unwanted
mutual inductances in these systems may produce unex­
pected and sometimes peculiar characteristics. On the other
hand, the dedicated experimenter may enjoy leaving sim­
plicity out of his circuit limitations for this very reason.

91
Appearance

Is the new project a reproduction of an early day radio? If

so, the restrictions on appearance are rigid and unyielding.
Components, materials, and layout must conform exactly
with the original set. There is no receiver design phase at all,
though tooling up for construction may itself prove to be a
major challenge.

The construction of a replica may present a different situ­

ation. The term replica has been applied to a wide range of
possibilities; near reproduction quality is at one extreme
while traditional external styling of a modem internal system
is at the other. In the special case of crystal set construction,
the tendency is to retain a traditional, i.e. early day, flavor for
the appearance of the front panel and major components. In
contrast, you may decide to try a new design and use modem
components throughout, with adaptations where necessary
for the special requirements of crystal sets. The most signifi­
cant adaptations are usually associated with the coils.

There is no unwritten law requiring the separation of modem

and early day styling, but some critics may object to such
cases as the use of a three inch diameter bakelite dial for
tuning the antenna, together with a broadcast band calibrated
slide rule dial assembly for the detector circuit. The old ad­
age about beauty being in the eye of the beholder can be
applied here, unless conformity with the past is an important
factor in the design.

92
Antenna Compatibility

The antenna circuit depends upon the type of antenna that is

available when optimum performance is wanted. The impor­
tance of antenna tuning has been repeatedly emphasized for
that condition when a tunable antenna is used. A medium to
long length antenna requires a variable capacitor in series
with the antenna and antenna coil. A very short antenna re­
quires that the capacitor be connected in parallel with the
coil. In some transitional cases it may be necessary to pro­
vide switching to accommodate both possibilities. See the
circuit shown in Fig. 3-2.

v
L 4

2 1
2i?1
4
C,

Figure 3-2: Antenna tuner switching ciruit for moderately

long antennas.

93
A tunable antenna may be impractical in some cases, such as
on a small lot or in an apartment building. A wire strung
around a room is not tunable. In these cases, a different ap­
proach to receiver design is needed. Consider adding only
one more tuned circuit ahead of the detector. Such a set
usually tunes rather broadly, but sensitivity is good. This is a
necessity for the low level signals that the antenna produces.

Design Options for Applications

The design criteria discussed earilier determine the circuitry

which can be used for a specified application. General
guidelines are helpful in making a selection, and these are
developed in following paragraphs. It is convenient to cate­
gorize crystal sets in order to do this. There are several rea­
sonable approaches to categorization, and a two-fold method
is used here.

First, there is a large class of receivers that depend primarily

upon variable capacitor tuned circuits for frequency selec­
tion. These are easily discussed, and it is possible to make
some general statements about their relative performance.
This class is presented in order of increasing complex­
ity—the number of tuned circuits present—which is an
important consideration when a simplicity compromise is
needed. The variable capacitors in these tuned circuits can
be replaced by fixed capacitors, with variable inductances
(variometers) used to do the tuning. Alternatively, a mixture
of variable capacitors and inductances can be used. Gener­
ally, these systems can be classified with those above.
The rejector wave trap is treated as a special situation.

Specialized variable inductance systems form a second class

of circuit choices. Their operating characteristics and degree

94
of performance are highly dependent upon the way variabil­
ity is built into the inductors. The number of tuned circuits is
not always as important as the particular design used for the
variable inductance and the manner in which they are used in
the system. These are best treated on a case by case basis.

Single Tuned Receivers

These simple systems use one tuned circuit, as shown in

Figure 3-3. Component values are for broadcast band appli­
cations, but these can be changed for other frequencies.

The circuit in Figure 3-3 A tunes very broadly unless the an­
tenna is deliberately shortened to just a few feet in length,
and it seldom gives satisfactory results. One way of improv­
ing the poor selectivity is to use a large loop antenna. A six
inch diameter or larger solenoid coil can also be used as a
loop antenna if the receiver is close to a local station. Selec­
tivity is good, but the received signal may be very faint. In
any of the systems described here some improvement may be
possible by using either a tapped coil or coil with a slider.
Another alternative is to move the variable capacitor to the
antenna circuit. This is often an improvement when a long
antenna can be used.

Two coils are used in the Figures 3-3B-D and these circuits
have a greater range of flexibility. By preadjusting the mu­
tual inductance or by building variocoupling into the receiver
coupling can be adjusted. Figure 3-3B works best with a
long antenna. The set of figure 3-3C works with a short an­
tenna, and also allows a longer antenna combined with
variocoupling .

95
V
4 4

^7 2
B

V V
4 -4

1
1U 2 D
c

Figure 3-3: Receivers with one tuned circuit.

Coil and Variable Capacitor Notes:

Untuned antenna coil wound with few turns for low

inductance.

Broadcast band coil. Typical inductance is 250ph.

Variable capacitor with nominal rating of 350 to 365pf.

96
One of the reasons single tuned circuits with a tuned detector
show poor selectivity is due to the energy drawn directly
from the resonant circuit to drive the earphones. This de­
creases the Q of the tuned LC circuit. The circuit in figure 3-
3D can be an exception to this situation, if loose coupling is
maintained between the detector coil and the tuned antenna.

Another example using an untuned detector is shown in Fig.

3-4C. Here a wave trap is coupled to the detector coil, which
is loosely coupled to the antenna coil. The net action of the
wave trap is to reject energy at the tuned frequency to the rest
of the system, and this energy is then available at the detec­
tor, while signals at other frequencies are non-resonant in the
system and have low amplitudes. Although the trap itself
performs as a rejector, the system as a whole accepts only the
desired signal. This approach has sometimes been called ab­
sorption tuning. Both mutual inductances can be established
by variocoupling to make this receiver a versatile and re­
markably efficient system. It can be expected to perform
comparably with some receivers employing two tuned
circuits which are described in the next section. This is also
true of the system which results when the wave trap is in­
ductively coupled to the antenna coil rather than the detector
circuit. Experiments are recommended to see which is best
for a given installation.

Figures 3-3B through 3-3D show one or more ungrounded

(floating) circuits. Locations 1 and 2 in 3-3B, C, and D may
to connected together and to ground on an experimental ba-
sis. Usually this improves sensitivity and decreases
selectivity.

Capacitive coupling may be substituted for inductive cou­

pling in any of these systems. However, there seems to be

97
no particular advantage in using this type of coupling, and at
least one early day reference indicated that broader tuning
would result. Also, additional components are necessary.
The same is true of more elaborate types of coupling circuits,
which are discussed in electronic engineering texts. They
require even more components than are needed for capacitive
coupling. Nevertheless, experimentalists may find it
interesting to investigate these possibilities.

The variable capacitor in circuits such as those of Fig. 3-3

can be replaced by a fixed unit, with tuning accomplished by
a variometer type of tuning coil. This method was widely
used in equipment built prior to the mid 1920s. Except for
reproductions and replicas, there is no real need for this ap­
proach.

All of the systems with tuned antennas show the variable ca­
pacitor on the ground side of the antenna coil, which is good
design practice. However, the broader tuning versions of the
sets are so nonselective that hand capacity is not important.
In such cases, the variable capacitor can be placed on the
antenna side of the coil if desired. The change in resonant
frequency caused by hand capacitance is not enough to pro­
duce a substantial change in signal level.

Most of the receivers in Fig. 3-3 use a minimum of parts,

making them relatively simple to build. This can also lead to
compact packaging as well as cost savings. Sometimes this
becomes especially important when early day components
are used. One variable capacitor is easier to find than a
matched pair, for example. Tuning these sets is easy, and a
beginner should try some of them just to get an idea of the
limitations on crystal set performance.

98
Receivers with Two Tuned Circuits

Some receivers with two tuned circuits appear in Figure 3-4,

and several are upgraded versions of the single tuned circuits
discussed in the previous section. The set of Figure 3-4A
usually shows a marked improvement in selectivity over
those of Figures 3-3 A and C, when variocoupling is also
available. It is still a relatively simple set to build and oper­
ate, although it is intended only for a long antenna. Figure 3-
4B is the very short antenna alternative. These sets can be
effectively combined in one receiver using the switching cir­
cuit in Fig. 3-2. Absorption tuning is used in Figures 3-4C
and D. Improvement may be relatively small, however,
since power loss in the earphones lowers the Q of the detec­
tor circuit.

V V
4
4 4

1L T
A 6
P cl§&
<=>2 9>

V
4

2
°1
Q3

£=4
P 1
i° 2

If P
i

D E

Figure 3-4: Receivers with two tuned circuits.

Fig. 3-4C works with a very short antenna when the antenna
coil is tuned. Fig. 3-4D may be a better choice for an un-

99
tuned antenna with intermediate length. The coils in Fig. 3-
4E are arranged so that energy received from the antenna is
transferred from the low inductance antenna coil, through the
tuned intermediate circuit, to the tuned detector. There is no
significant amount of coupling between the antenna and de­
tector coils. However, the benefit of antenna tuning is not
utilized here, so performance is limited and usage is
suggested only where a tunable antenna cannot be employed.

The tapped coil and slider coil options, switching to tune

either a long or short antenna, and the ground connection
possibilities, all discussed for single tuned receivers, are also
applicable here. Variocoupling is another desirable feature.
Adding a tuned circuit results in the type of receiver covered
in the next section.

Receivers with Three Tuned Circuits

Receivers with three tuned circuits were employed well be­

fore World War I, and were recommended as a desirable
combination of three general methods of tuning. These
systems show the high degree of sophistication achieved
even before World War I. Most capacitors and inductances
in these systems are variable, and this permits selecting op­
timum values of all components for best operation. There
are disadvantages nonetheless. These are very complicated
systems, so the operator must be highly skilled, familiar with
the receiver, and acquainted with local receiving conditions.
Also, the coils are not as efficient as might be supposed, be­
cause the unused open-ended portions lower the Q, due to the
dead end effect. The latter is caused by self-induced voltages
in these unused windings, with consequent losses.

100
 V V
4 4
i
A

Figure 3-5: Receivers with three tuned circuits.

The arrangement in Fig. 3-5A is a much simplified version

of the early day three tuned circuit receiver. It has two tun­
ing capacitors in the antenna circuit, plus a tuned detector.
The two antenna capacitors are strongly interactive. That is,
changing the tuning of one makes it necessary to fully retune
the other. The best combination of dial settings depends
upon the length of the antenna and the signal frequency.
This receiver is not easy to operate at its best, even though it
is much less complicated than the early day version.

Fig. 3-5B is an improved version of the double tuned re­

ceiver in Fig. 3-4A, with absorption tuning added at the de­
tector coil to increase selectivity. Sometimes a simple
rearrangement of the coils will make a difference in
performance.

101
The overall efficiency can be very high using three tuned cir­
cuits. Excellent selectivity is possible without the drastic
sacrifice of sensitivity that is usually necessary with simpler
systems. On the other hand, a relatively large number of
components is required, as well as additional layout and
construction work. Also, tuning three variable capacitors can
become tedious, particularly with the degree of selectivity
which can be reached with such receivers.

The antenna tuner circuit cannot be expected to track the

other tuned circuits across the entire broadcast band, but with
some sets the other two variable capacitors can be ganged
together for simpler tuning. Alternatives for making com­
mon connections or connections to ground are reduced when
a conventional two gang tuning capacitor is used, because of
the metal rotor shaft common to both units.

Detector circuits generally track well with an absorption

tuned circuit or energy transfer circuit, when a fixed tap is
used on the detector coil. The two circuits can be adjusted
for tracking by means of the trimmer capacitors usually pro­
vided with the two gang unit, or separate trimmers can be
installed. Tracking may be poor when a multi-tapped high Q
detector coil is used. This occurs because the wiring to the
taps introduces additional stray capacitances around the de­
tector coil, changing its tuning characteristics. The wiring
needs to be adjusted only if a large change is made in the
tuned frequency. A slider may be used instead of multiple
taps with less effect upon tracking, and it is an attractive
choice for this application.

High Q systems have another disadvantage that becomes

very noticeable when large diameter air core coils are used.
Stray capacitances and unwanted mutual inductances become

102
so large that circuit characteristics are altered, usually in un­
predictable and undesirable ways. These effects can become
important when three inch or larger diameter coils are used,
and they depend upon the size of other parts and the physical
layout, as well as the particular circuit that is used. Ferrite
core coils, with their combination of high Q and small size,
can be used to advantage in these systems, because the un­
wanted side effects are greatly reduced. Unwanted mutual
inductances also can be reduced by using closed field type
binocular coils.

Variable Inductance Tuned Receivers

Many crystal sets, especially the early day models, tune by

some method of changing inductance rather than by chang­
ing capacitance. Inductance is changed either by a series of
taps from the coil to a switch, by a sliding conductive contact
along the coil, or by a variomoter. These possibilities were
mentioned earlier. Combined variable inductance and vari­
able capacitance tuning is also used in some cases. A great
number of circuit combinations are possible, and in some
cases performance may be more dependent upon the variable
inductance design and circuitry than upon the number of
tuned circuits. It is impractical to separately discuss all of
these combinations, and they can be considered as variations
of the receiver designs discussed earlier.

Wave Trap Applications

Circuits can be added to receivers in any of the categories

covered earlier, either for overall improvement, or for a spe­
cific application. Acceptor wave traps were discussed
earlier, because of their frequent use for absorption tuning in
capacitance tuned circuits. However, rejector wave traps are

103
used quite differently. One or more rejector wave traps can
be used with any receiver, but this is considered a special
application. A rejector trap affects receiver performance in
the portion of the band near its tuned frequency rather than
over the band as a whole. Nevertheless, the only practical
use is to preset it and leave it. The receiver tunes over most
of its frequency range as if the trap were not present.

As stated in Chapter two, the most commonly chosen loca­

tion for a rejector wave trap is in the antenna circuit.
Rejection is at least as efficient in the antenna circuit as at
any other location, and there is less disruption of normal
tuning at nearby frequencies. The opinion of some crystal
set enthusiasts is that it is better to improve overall receiver
selectivity by other methods than to use a wave trap to
minimize an interference problem. Nevertheless, a rejector
trap may be worthy of consideration for an unusual special
case, such as one very powerful local station which is a ma­
jor cause of interference. Wave traps of either type can be
added to the circuits described.

104
BIBLIOGRAPHY

Anderson, John E., Arthur C.C. Mills, and Elmer H. Lewis,

Henley's 222 Radio Circuit Designs. Korman W. Henley
Publishing Co., New York, 1923, p.67. Reprinted by
Lindsay Publications, Inc., Bradley IL, 1989.

Anderson, Phil, The Miller '595' Tuner Revisited. The Xtal

Set Society Newsletter. Vol. 2, No. 5, May. 1993, p. 1.

Bucher, Elmer E., Practical Wireless Telegraphy. Wireless

Press, Inc., New York, 1917, p. 144.
Dynner, Eugene, An Exceptional Radio Receiver. Radio
Amateur News, Vol.l, No.2, Aug. 1910, p. 63.

Gardner, Leonard U., W2QBC, 458 Two Mile Greek Rd.,

Tonawanda, NY 14150, personal communications, Jan. 31,
1991 and Jan. 31, 1993.

Harrison, Arthur, Constructing Wireless-Era Receivers from

Authentic Components. Antique Radio Classified, Vol. 9,
No. 11, Nov. 1992, p.18.

Hayward, James, At Last, a Portable Crystal Set. The Xtal

Set Society Newsletter, Vol. 4, No. 1, Jan. 1, 1924, p.2.

Horvath, Joe, Uncle Joe's No. 4 Super Selective Single Dial

Control Crystal Set. California Antique Radio Gazette, Vol.
11, No. 2, May 1986, p. 14.

Horvath, Joe, Coil Data for Uncle Joe's Super Selective

Tuner or Crystal Set. Op. Cit., Vol. 12, No. 1, Feb. 1987, p.
23.

105
Kinzie, P.A., Some Experiments with Crystal Sets Using
High O Coils. Arizona Antique Radio Club News, Vol. 7,
No. 1, Spring 1990, p. 12.

Pierce, George W., Principles of Wireless Telegraphy.

McGraw-Hill Book Co., Inc., New York, 1910, p. 286.

Osterhoudt, Elmer G., MRL 13 Crystal Set Circuits.

Handbook HB-25. Modem Radio Laboratories, 1947,
Minneapolis, MN.

Osterhoudt, Elmer G., MRL Detail Print No. DP22.

Osterhoudt, Elmer G., MRL Detail Print No. DP23.

Osterhoudt, Elmer G., MRL Detail Print No. DP32.

Strozier, Chuck, A Crystal Set. California Antique Radio

Gazette, Vol 15, No. 4, Nov. 1990, p. 18.

106
 w iu CHAPTER 4
CRYSTALS FOR DETECTORS
*
L1 O

Materials for Detector Applications

Q L3
CT3
Most active substances which have
been used in crystal sets are
crystalline materials which originated either in ore deposits
or as synthetic manufactured products. As usual, there are
exceptions to the rule, such as the fused silicon product of
the electric furnace, which can be non-crystalline. The in­
tention here is to discuss the practical application of these
materials rather than to precisely categorize. Therefore, if it
looks like a detector material, acts like a detector material,
and is used like a detector, it is included.

It is surprising how few unusual cases are included, using

this selection rule, in view of the lack of restrictions in the
statement. Two of these are electrolytic detectors, using lead
peroxide or zinc peroxide. They were used in the form of
compressed powder pellets. A few other materials depend at
least in part on the principle of imperfect contact rectifica­
tion. The carbon-steel detector qualifies here, the extreme
example being the "fox hole" radio of World War II. Its de­
tector was a piece of pencil lead resting across two parallel
razor blade edges.

There are some less clear-cut examples of active substances

showing imperfect contact behavior. Some samples of one
or two crystalline compounds have been reported to improve
in sensitivity as the surface becomes tarnished, when a cat-
whisker contact is used. One possible explanation involves
imperfect contact supplementing the usual solid state rectifi­
cation property. It should be noted, however, that this

107
behavior is very rare. Performance usually deteriorates as
the surface of a crystal oxidizes.

There are no firm rules regarding the composition and ap­

pearance of a substance which might establish it as a good
solid-state rectifier. There are only trends that suggest the
possibility. Early day investigators were hampered and
sometimes misled by this state of confusion; a satisfactory
theory did not exist to explain rectification. Early work was
entirely empirical, including the first demonstration of solid-
state rectification by Ferdinand Braun. Braun noted some of
the trends that led to good examples, being first to point out
that metal sulfides are noteworthy in this respect.

Later investigators noticed a certain degree of correlation

between rectification and the thermoelectric properties for
many intermetallic compounds, such as tellurium, which is
electrically conductive to some extent. The correlation was
so noticeable that for a time it was widely believed that
thermoelectric effects caused rectification due to the heat
presumably generated at the point contact. It required some
very difficult laboratory experimentation to disprove this
theory.

The appearance of a material sometimes gives a clue about

its performance, but this test is inconclusive. Most good de­
tectors have a shiny metallic appearance ranging from silvery
white through steel gray. Golden and coppery crystals also
occur. Nevertheless, a few detectors have a dull nonmetallic
appearance.

There are other material properties that have important ef­

fects on the preparation and use of an active substance as a
crystal detector. Heat sensitivity is a problem with some

108
 materials, including galena, one of the most used crystals.
Heat may either accelerate surface oxidation or alter the bulk
material, and this degrades sensitivity. The sensitivity of
galena and certain other crystals may be destroyed locally at
the point contact by static discharges or even by a very
strong signal, all because of the sensitivity to heat that is
generated at the contact. Mounting such crystals in a metal
base requires special care, and this will be discussed in the
next section.

Most crystal surfaces are all too easily damaged by careless

handling. Your fingers should not touch an active surface. It
may be possible to clean a freshly contaminated crystal with
methyl ethyl ketone (PEK) or acetone, but no great degree of
success has been reported in this respect. One exception,
zincite has been reported to be cleanable with carbon disul­
fide after becoming contaminated. On a longer term basis,
most materials exhibit a progressive deterioration simply
through exposure to air. Tarnish caused by surface oxidation
is the causative factor. The rate of oxidation varies with
composition. Some types of crystals need to be replaced
! after a few months. With others the useful lifetime may be
3 measured in years. In any case, heat speeds up the deteriora­
tion process.

Typical detector materials are relatively hard and brittle.

Even so, just the repeated adjustment of a catwhisker may
eventually mar the surface, resulting in decreased perform­
ance. The use of sharp pointed contacts or the broad end
contact of a second crystal can easily destroy the surface as
far as sensitivity is concerned. Extreme care must be taken
while making adjustments to these types of connections.

109
There are almost no exceptions to the rule that attempts to re­
surface a crystal, by polishing, sanding, or grinding, will
destroy it. The noteworthy exception is molybdenite. This
compound can actually be restored to high sensitivity by
sanding the exposed surface. In practically all other cases,
the only alternative to replacement is to split the crystal and
remount it with a fresh surface exposed. Even this procedure
does not work if the interior has been spoiled by overheating.

Up to this point, materials properties have been discussed

only in a generalized way. Tens of thousands of different
compositions and combinations of materials have individual
characteristics that have been tested for rectification over the
years, and most results were never made public. It can be
assumed that the reason for such lack of information is sim­
ply that performance was poor, so there was no incentive to
report the results.

Contacting Methods

Point Contact Junctions

The first successful point contact method for practical crystal

detectors depended upon a stiff, pointed wire, or even a nee­
dle or pointed rod. Contact pressure varied, depending upon
the crystal. .-In some cases this provided a fairly stable
detector for"rugged applications onboard ships and in army
field equipment. The most sensitive and yet practical mate­
rials with this type of contact were carborundum (silicon
carbide), cerussite (lead carbonate), and molybdenite
(molybdenum disulfide). Silicon was also used to a degree.
This was the fused silicon electric furnace product, not to­
day's highly purified, doped silicon crystals.

110
The light wire contact method, which includes the very deli­
cate catwhisker just touching the active substance surface,
evolved later. The catwhisker contact was patented by
Greenleaf Whittier Pickard, one of the inventors of the crys­
tal detector itself, but only after several years had passed
following his first patent. Pickard later remarked that the
catwhisker must have been independently invented and used
by others before he filed for his patent.

Initially, there was a lack of interest in the catwhisker, be­

cause it is so susceptible to vibration and shock. However,
its performance in a benign environment is unequaled by
other methods, when a crystal of the proper composition is
used. It became the most used contacting method when the
advent of broadcast radio provided a demand for high sensi­
tivity without the pressing need for reception in rough
environments. This brought about a somewhat different
choice of crystals, with compositions for use in the home.
The materials with the highest sensitivity were steel galena,
galena, iron pyrites, and fused silicon. At present, the ger­
manium diode is the fixed crystal substitute for the galena
materials. It has equal or nearly equal sensitivity, and there
is no requirement for adjustment.

A comparison of the sensitivity and performance of pointed

wire contact versus catwhisker contact is difficult to evaluate
when there is little knowledge of which method was used
with crystals. For example, textbooks published before
about 1912 usually discuss sensitivity in terms of pointed
wire or rod contact, and list the most sensitive crystal as ce-
russite or molybdenite, with carborundum also included. In
contrast, once the catwhisker came into use, textbooks listed
galena and iron pyrites as sensitive. The composition of the
catwhisker wire itself also appears to make some difference

ill
in performance. However, there is a lack of agreement con­
cerning which metal is best, because in some cases the
choice of crystal affects the choice of catwhisker metal,
which has complicated the acquisition of data.

There is another contact method, not previously mentioned,

and which also preceded the catwhisker. This was the use of
a second active substance, usually in the form of a pointed
crystal, held in contact with the surface of the first under
moderate pressure. The use of two crystals for detection was
another Pickard invention, and it is commonly referred to as
a Perikon. This is a generalization of Pickard's original name
for a specific pair of materials. Like a pointed rod under
pressure, a Perikon detector was suitable for rugged envi­
ronments and was used to some degree for shipboard
installations. However its sensitivity did not equal that of a
catwhisker.

Other Contacting Methods

In modem fixed diodes no fully satisfactory substitute exists

for a point contact junction, established either by mechanical
pressure, or by a delicate spot weld. Results with other
methods have been partially successful at best. The earliest
carborundum detector is the classic example. It consisted of
a relatively long carborundum crystal wrapped with copper
wire at each end. The rejection rate during the first produc­
tion run was on the order of 90 percent. It was later
concluded that rectification in the better detectors was due to
a pronounced bulk properties change along the length of the
crystal. A relatively uniform crystal detected poorly.

A few other active substances have been used by mounting

them between two plates under pressure. Different plate

112
 *■

metals were sometimes used to improve detection. The

shape of the contacts was a factor in some cases; at least two
patents were issued in the 1900-1910 period which described
a cylinder of metal contacting the active surface of the mate­
rial. None of these methods equaled the sensitivity of a fine
wire point contact junction.

A closer approach to satisfactory operation was obtained

through the use of either a powder or a brush of many fine
wires as a substitute for a catwhisker. Enclosed, even scaled
cartridges, were sold during the first years of the broadcast
era, beginning in the early 1920's. Usually the sales pitch
emphasized that many separate contacts eliminated the
problem of a sudden loss of contact — which was character­
istic of a disturbed catwhisker — as well as the frustrating
search for a sensitive spot. Some of these products provided
for user adjustment "just in case."

The Base Connection

The best form of base connection for point contact detectors

consists of a good wide area electrical contact. This can be
produced by casting the active substance, partially immersed,
in a short metal cup with some low melting point metal. Ac­
ceptable performance is usually possible if the specimen is
placed in a hollow holder and a soft metal wool or foil is
tightly packed around it. Fine steel wool has enough of a
packing characteristic to be used in the way.

113
Application Technique

Choosing a Crystal

At present the selection of a crystal for new construction

projects is relatively simple, as far as choosing from existing
products is concerned. These include solid state diodes as
well a premounted crystals. The 1N34 germanium diode
gives good performance along with the advantages of con­
stant sensitivity and no requirement for adjustments. Other
diodes that have been used and recommended include the
1N21, 1N22, 1N60, 1N66, 1N69, 1N128, 1N294, and
1N295. Silicon diodes work well but have no particular ad­
vantages, and they require a small amount of bias voltage for
best results.

The property of constant sensitivity is highly desirable when

the object is to evaluate other components or circuit changes.
There is no added variable to confuse the results. In other
words, if circuit B works better that A, there is no likelihood
that change was caused by the detector instead of the circuit.
This may occur when using a catwhisker crystal. However,
when sensitivity is all-important, a galena crystal works
least as well as a 1N34, and occasionally even better. Some
steel galena crystals have the same level of sensitivity. Iron
pyrites are not quite as sensitive as the galena types for the
broadcast band, but they are the preferred choice for short
wave receivers, Silicon is satisfactory for the broadcast
band, but carborundum is definitely less sensitive, even with
the proper bias voltage applied.

The picture is quite different for the case of investigating

crystals themselves. A study of modem synthetics, which
were unavailable to early day experimenters, is an interesting

114
field for those who like to be among the first to try some­
thing new. There is also the possibility of finding good
catwhisker crystals among the materials which were origi­
nally used with heavy contact methods and were discarded.
Also, new sources of ore have been found since the early
wireless period, and they may have better properties than
those previously used. In short, there are a large number of
possible sources to choose from, but locating the specimens
can be a limitation.

Different limitations are associated with restoration projects

or with construction of a reproduction. This is because most
equipment manufactured before the end of World War I did
not use a catwhisker crystal for detection. Therefore, it
would be inappropriate to choose a galena crystal and cat­
whisker for such a project without supporting
documentation, even though the method was well known by
that time. The only certain way to make the right choice is to
determine what was used in the specific receiver.

Some general guidelines are available when specific infor­

mation cannot be found and typical usage is an acceptable
criterion. Pre-World War I American Marconi receivers
used carborundum crystals for several years, along with
heavy pointed contacts under considerable pressure. The
parent Marconi company in England adapted the same type
of detector after retiring earlier detection methods. Some­
what later, both carborundum and cerussite crystals were
used in dual holders that allowed an operator to switch from
one to the other. Silicon was used on detector stands sold by
the Wireless Specialties Products Company. These were in­
stalled in a number of different receivers. Molybdenite was
also available, and there were Perikon type combinations

115
with two different crystals also in use. All of these were of
the heavy contact junction type.

Wireless operators on ships sometimes carried their own fa­

vorite crystals with them, regardless of company policy
discouraging that practice. Amateurs were noted for skill
and interest in experimenting. Therefore, it is likely that a
wide variety of materials actually received at least some us­
age. As stated earlier, the catwhisker technique was also
known, although the most likely users were amateurs and
shore station operators in a relatively quiet environment.
When evidence of such usage on a specific set exists, it is
reasonable to restore it for the same less-than-typical choice
of detector.

There were factors which affected the choice of a crystal

during the early wireless era which are not obvious under
more modem conditions. Some understanding of these is
helpful when working with early day designs. First, carbo­
i
rundum was often the crystal of choice because of its
ruggednesss and stability, even though it was less sensitive
that the other commonly used materials. Its tolerance of heat
and high currents allowed it to be used where current surges
that damaged materials more susceptible to heating were apt
to occur. Static discharges from the long, high, shipboard
antennas then in use were one source of surges. Further, by
necessity, shipboard transmitters and receivers either used
the same antenna or closely adjacent lengths of wire, and
were located close together. In spite of precautions to avoid
high currents through the detector, there was a problem with
the less heat resistant materials. In fact, a strong local signal
from another ship or nearby shore station was sufficient to
cause damage in some cases. Carborundum could survive
under such conditions.

116
The hardness associated with carborundum was another ad­
vantage. A stiff wire, needle, or pointed rod could be held
against the surface of the crystal under considerable pressure
without destroying a sensitive spot. High pressure greatly
decreased the chances of vibration or shock causing contact
point movement and loss of sensitivity. Other materials were
not as tolerant of contact point pressure, and thus they were
more prone to being knocked out of adjustment.

The fused silicon electric furnace product then available was

considered to be generally satisfactory, although somewhat
less rugged than carborundum. Higher sensitivity was a
compensating factor. Cerussite and molybdenite are softer
than either carborundum or silicon, so lighter contact pres­
sure was necessary, leading to more problems with vibration
and shock. They are also more sensitive to heating, but when
these properties could be accepted, the resulting sensitivity
was higher than either carborundum or silicon. Galena was
well down on the list of crystals because of its poor heat
sensitivity and rapid surface oxidation characteristics. When
used with a heavy contact it was not as sensitive as cerussite
or molybdenite. Other minerals which were tried had similar
disadvantages and were little used.

117
BIBLIOGRAPHY

Constable, Anthony, Early Wireless. Midas Books, 12 Dene

Way, Speldhurst, Tunbridge Wells, Kent, England, 1980.

Gibbons, Robert C., (Editor), Woldman's Engineering

Alloys. Sixth Edition. American Society for Metal, Metals
Park, Ohio, 1979.

Knoll, Max, Materials and Processes of Electron Devices.

Springer-Verlag, Berlin, 1959.

Kohl, Walter H., Handbook of Materials and Techniques for

Vacuum Devices. Reinhold Publishing Corp., New York,
1957, p. 358.

Osterhoudt, Elmer G., Crystal Detectors. MRL Handbook

HB-3, Modem Radio Laboratories, Minneapolis, MN 1954.

Osterhoudt, Elmer G., Facts for Crystal Experimenters. MRL

Handbook H9-1Q. Modem Radio Laboratories, Minneapolis,
MN, 1954.

Sievers, Maurice L., Crystal Clear Volume I. The Vestal

Press, Ltd., Vestal, New York, 1991.

118
ABOUT THE AUTHOR

P.A. (Phil) Kinzie earned a Master's Degree in Engineering

at UCLA, and worked with aircraft ground and flight test
instrumentation for many years, with later similar
employment in the space program. More recently he was
registered as a Control Systems Engineer in California, with
experience in oil refinery and similar industrial projects.

Although not currently in engineering, he has continued a

long-time interest in crystal sets that began at the age of nine
or ten. At that time he lost his enthusiasm for dismantling
broken clocks and watches, and turned his attention to old
radios. He dissected these to extremes that included
unwinding the paper-foil capacitors to see what was inside.

Destruction is not usually the recommended way of entering

a new field, but in his case he eventually graduated to a
better pastime of building crystal sets and one tubers, a
hobby that he still continues when time permits. This book
reflects his effort to analyze and improve upon the special
technology needed for design and construction of crystal
sets.

119
I

:
 V THE XTAL SET SOCIETY

The Xtal Set Society Newsletter, bi-monthly, one

year subscription, (postage is included, no additional
shipping charge) S12.95
International subscriptions please remit US$19.00
Canadians please remit US US$14.00

Volume I & II of the Society Newsletter, twelve issues,

May, 1991 through May 1993. $ 19.95

The Crystal Set Handbook and Volume III of the

Society Newsletter, three issues, ending Nov. 1993. $ 11.95

Volume IV of the Society Newsletter, six issues,

ending November 1994. $ 10.95

Crystal Sets: Volume V of the Society Newsletter,

six issues, ending November 1995. $ 10.95

Crystal Set Projects: 15 Radio Projects You Can Build

New book! Written by members of the Society, 1997 $ 14.95

Crystal Set Building and More: Volume 6 & 7 of the Society

Newsletter, 12 issues ending November 1997 $ 15.95

Crystal Set hoopers, a Three Tuber & More: Volume 8

of the Society newsletter, 6 issues ending November 1998 $ 14.95

Shipping and handling on book orders $ 3.50

The Xtal Set Society

www.midnightscience.com
e-mail: xtalset@midnightscience.com
1-800-927-1771 VISA/MC accepted

121
 Crystal Set Projects: 15 Radio
: Crystal Sft
Projects Projects You Can Build
Crystal Set Projects is a collection of radio
T projects that won a recent building contest run by
i the Society. Thanks to everyone who par­
ticipated. If you missed out, don’t worry. We’re
doing another! Included are step-by-step
instructions so you can build and design your
w 1 * own Xtal sets no matter what your background in
electronics. The projects are of various designs
and difficulty so that everyone can learn
something. “Crystal set circuits have been regularly published in the
hobby press since the dawn of radio. But the Xtal Set Society's new
project book covers more electronic and mechanical variations of this
ever-popular receiver than I have ever seen in one place. You'll find hi
tech and low tech versions, vintage and modem versions, sets with loop
antennas and short-wave bands. The book is as much fun to read as to
build from, but I defy anyone to browse this infectious publication for
long without experiencing an uncontrollable urge to break out the
soldering iron and double-colton-covered wire!” —Marc Ellis, Editor,
The Old-Timer's Bulletin of the AWA; Contributing Editor, Antique
Radio, Popular Electronics magazine. Get yourself a copy! 6x9
paperback, 160 pages, $14.95.

The Xtal Set Society is dedicated to once again building and

experimenting with radio electronics, often—but not always—through
the use of the crystal set, the basis for most modem day radio apparatus.
The Society newsletter helps support our goal of producing excellent
quality technical books that encourage learning and building. To join the
society and receive one year (6 issues) of the bi-monthly newsletter,
remit $12.95 to The Xtal Set Society. Canadians, please remit US
$14.00. Outside the US and Canada please remit US $19.00.

The Xtal Set Society

1-800-927-1771 VISA/MC
e-mail: xtalset@midnightscience.com
www.midnightscience.com
utgZs''
jv \f
m is&J

122
 :

■
 This book chronicles the fascinat­
ing history and development of
the crystal detector including the
reasons for its brief dominance of
the radio market after the turn of
the century. Starting with the dis­
covery of solid-state rectification
and then through the develop­
ment of the crystal detector the
reader learns about great inventors
such as Pickard, Braun, Dun-
woody, and others. Radio funda-
^ mentals such as antennas, ground,
pl1gl|hFiing{|fi^t6ction, cir­
cuits, and detection are covered for
the beginner. The unending com­
promise between selectivity and sensitivity is discussed for the "
crystal set d&srgner. Advanced topics such as the use of multi-
tuned eirff|^lind wave traps follow for the more experienced ^
experiment‘d

The crystal set embodies many of the basic concepts upon

which modern radio electronics is based. By grasping these
basics in their rarest form, the crystal set, the reader can build a
powerful foundation for additional learning and designing.

Marconi transmits Picka

successful 'wireless' silic
1895
r N
-
Y
1874
ISBN 1-887736-05-0 oliotheek Ned. Vel
Braun discovers unilateral
conduction (rectification)

The Xtal Set Society

www.midnighlscienee.coni 9 781887 736053
1-800-927-1771